Open Collections

UBC Faculty Research and Publications

Survival benefit of a low ratio of visceral to subcutaneous adipose tissue depends on LDL clearance versus… Lee, Joseph G H; Genga, Kelly R; Pisitsak, Chawika; Boyd, John H; Leung, Alex K K; Russell, James A; Walley, Keith R Mar 6, 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13054_2018_Article_1985.pdf [ 760.3kB ]
JSON: 52383-1.0364170.json
JSON-LD: 52383-1.0364170-ld.json
RDF/XML (Pretty): 52383-1.0364170-rdf.xml
RDF/JSON: 52383-1.0364170-rdf.json
Turtle: 52383-1.0364170-turtle.txt
N-Triples: 52383-1.0364170-rdf-ntriples.txt
Original Record: 52383-1.0364170-source.json
Full Text

Full Text

RESEARCH Open AccessSurvival benefit of a low ratio of visceral tosubcutaneous adipose tissue depends onLDL clearance versus production in sepsisJoseph G. H. Lee1, Kelly R. Genga1, Chawika Pisitsak1,2, John H. Boyd1, Alex K. K. Leung1, James A. Russell1and Keith R. Walley1*AbstractBackground: Patients with sepsis with a high ratio of visceral adipose tissue (VAT) to subcutaneous adipose tissue(SAT) have increased mortality. Our goal was to investigate the mechanism of this effect, noting that low LDL levelsare also associated with increased sepsis mortality. Accordingly we tested for association between VAT/SAT, low-density lipoprotein (LDL) levels, and mortality. Then we examined the effect of statin treatment, which decreasesLDL production, and the effect of PCSK9 genotype, which increases LDL clearance.Methods: We performed retrospective analysis of a cohort of patients with sepsis from a tertiary care adultintensive care unit in Vancouver, Canada, who underwent abdominal computed tomography (CT) (n = 75) forclinical reasons. We compared LDL levels in patients with sepsis according to high versus low VAT/SAT and 90-daysurvival. We next examined the effects of statin therapy and PCSK9 loss-of-function genotype on survival.Results: Patients with a low VAT/SAT had increased 90-day survival and were relatively protected against low LDLlevels in sepsis compared to high VAT/SAT. Statin treatment abrogated the beneficial effects of low VAT/SAT;eliminating the difference in LDL levels and survival between patients with low and high VAT/SAT. PSCK9 loss-of-function genotype similarly eliminated the increased LDL levels in low VAT/SAT patients but, in contrast, increasedthe survival advantage of low VAT/SAT compared to high VAT/SAT.Conclusions: Low LDL levels per se are not simply associated with decreased sepsis survival because lowering LDLlevels by inhibiting LDL production (statin treatment) is associated with adverse outcomes, while increased LDLclearance (PCSK9 loss-of-function genotype) is associated with improved outcomes in patients with low VAT/SAT.Keywords: Sepsis, Visceral abdominal fat, PCSK9 genotype, LDL-cholesterolBackgroundA number of studies report that obesity, as indicated byhigh body mass index (BMI), is paradoxically associatedwith improved survival during sepsis [1] although thisassociation has not been uniformly observed [2, 3]. Dis-crepancies between studies may be partly explained bythe recent discovery that patients with a high ratio of ab-dominal visceral adipose tissue to subcutaneous adiposetissue (VAT/SAT) have increased mortality from sepsiscompared to patients with low VAT/SAT [4] at any levelof BMI [4, 5]. Thus, the distribution of adipose tissue inobesity appears important in sepsis. The mechanism ofthis VAT/SAT effect is not understood.We postulated that lipoproteins may be involved since re-lationships between adipose tissue quantity and lipoproteinlevels have been reported in non-septic states [6]. Low LDLlevels are a risk factor for sepsis survival and also a risk fac-tor for incidence of sepsis [7]. Since pathogen lipids are se-questered within lipoproteins such as LDL during sepsis,decreased production of LDL leading to low LDL levels maydecrease the buffering capacity of LDL for pathogen lipids[8, 9] and allow unbound pathogen lipid to trigger a greaterinflammatory response leading to adverse clinical outcomes.* Correspondence: Keith.Walley@hli.ubc.ca1Centre for Heart Lung Innovation, University of British Columbia, Vancouver,BC, CanadaFull list of author information is available at the end of the article© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (, which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver( applies to the data made available in this article, unless otherwise stated.Lee et al. Critical Care  (2018) 22:58 contrast, increased clearance of LDL during sepsis (forexample, by loss-of-function genotypes of PCSK9) may bebeneficial [10, 11] because pathogen lipids within LDL parti-cles are cleared along with LDL more rapidly [10]. Thus,considering only static serum LDL levels may be too sim-plistic. Instead, we suggest that it may be necessary to con-sider the underlying balance of LDL production versus LDLclearance mechanisms when considering prognostic predic-tion of survival according to LDL levels in sepsis.To understand potential mechanisms that explain whyincreased VAT/SAT ratio is harmful in sepsis, we com-pared two known pathways affecting LDL production andLDL clearance. Statin therapy lowers LDL in the circula-tion primarily by inhibition of LDL production in hepato-cytes [12–14]. Statin treatment is common and thereforesufficiently prevalent in patients presenting with sepsis tohospital emergency departments to serve as a test of theeffect of decreased LDL production on LDL levels andsurvival in sepsis. PCSK9 inhibitors, which increase LDLclearance, have just been approved to treat certain types ofhyperlipidemia [15] but their use is not yet sufficientlyprevalent to examine their effect on increased LDL clear-ance in sepsis. We therefore chose the alternative strategyof measuring PCSK9 loss-of-function genotype, which isknown to increase LDL clearance. Thus, in patients withlow versus high VAT/SAT ratios, we examined the effectof statins to understand the effect of decreased LDL pro-duction, and examined the effect of PCSK9 loss-of-function genotype to understand the effect of increasedLDL clearance in patients with sepsis.MethodsPatientsWe conducted a retrospective analysis of a prospectivelyaccrued cohort of patients with sepsis enrolled from Janu-ary 2011 to July 2013 at St. Paul’s Hospital, Vancouver,Canada. Patients with suspected sepsis were identifiedwhen the Emergency Department physician activated theInstitutional Severe Sepsis Order Set. As these patientswere recruited prior to the new Sepsis-3 guidelines [16], weabided by the previous sepsis definition, which required pa-tients to have two or more systematic inflammatory re-sponse syndrome (SIRS) criteria and have known orsuspected infection [17]. Of these patients, we selectedthose who had an abdominal computed tomography (CT)scan for clinical reasons within 2 months prior to ICU ad-mission, during the hospital stay, or within 1 month afterdischarge. Statin use and intensity regimen (low, moderateor intense) were determined at the time of study inclusionaccording to the American College of Cardiology guidelinesfor statin therapy [18]. All patients provided written in-formed consent. This protocol was reviewed and approvedby the St. Paul’s Hospital/University of British Columbia re-search ethics board.Measurement of adipose tissue areaWe measured the visceral adipose tissue area (VAT) andsubcutaneous adipose tissue area (SAT) from the ab-dominal CT scans at three levels: between L2 and L3, L3and L4, and L4 and L5 as previously reported [4]. AllCT scan images were de-identified and then loaded into3D Slicer (, an open source softwarepackage for medical image visualization and analysis.VAT was identified by manually tracing a region ofinterest beneath the abdominal wall on the cross-sectional CT scan and adipose tissue area within this re-gion of interest was measured by determining the num-ber of pixels within a window width of − 190 to − 30Hounsfield units [19–21]. Similarly, SAT was identifiedby manually tracing the skin surface and abdominal wallon the cross-sectional CT scan and the same Hounsfieldunit window width was used to measure SAT. We calcu-lated area and used the average from the three levelsranging from L2 to L5 to reduce noise and errors. Thisanalysis was performed by the same individual for all CTscans while blinded to clinical outcomes.LDL measurementA 6-mL EDTA tube of blood was collected at the timeof initial blood culture and immediately placed on ice.Blood was spun at 1800*g for 12 min and total choles-terol (TC), triglyceride (TG) and high-density lipopro-tein (HDL)-cholesterol were measured in the plasmafraction on our hospital clinical laboratory ADVIA 1800Chemistry System (Siemens). LDL measurement was de-termined using the formula:LDL ¼ TC‐HDL‐TG=5:0:PCSK9 genotype measurementTotal genomic DNA for genotyping was extracted fromthe buffy coat fraction using a QIAGEN DNeasy Blood& Tissue Kit (QIAGEN 69506). PCSK9 genotyping wasperformed in all samples for three common PCSK9 mis-sense loss-of-function variants (minor allele frequency,≥0.5%): rs11591147 (R46L), rs11583680 (A53V), andrs562556 (I474V) and the gain-of-function variantrs505151 (G670E) using pre-validated TaqMan SNPGenotyping Assays (ThermoScientific 4,351,379). Assayswere run on a ViiA7 platform using the system’s existingSNP genotyping software and protocol. All alleles werecalled using the software clustering algorithm, with a100% success rate. For quality control, 10% of sampleswere randomly repeated for each SNP to ensure repro-ducibility. For simplicity in this genotype analysis weexcluded four patients who carried the rare PCSK9 gain-of-function variant.Lee et al. Critical Care  (2018) 22:58 Page 2 of 9Statistical analysisFor all comparisons we divided patients into those witha VAT/SAT below the median value (low VAT/SAT) andthose with a VAT/SAT above (high VAT/SAT) [4]. Wecompared baseline characteristics between low and highVAT/SAT groups using the t test for normally distrib-uted variables and the chi square test for proportions.LDL levels were not normally distributed so these dataare reported as median and interquartile range, and theMann-Whitney test was used for comparisons. Wetested for differences in 90-day survival curves using thelog-rank test. All analyses were conducted in SPSSchoosing p < 0.05 as significant.ResultsPatientsThere were 75 patients who fulfilled the inclusion cri-teria, of which 2 patients did not have an adequate qual-ity CT scan, leaving 73 patients for the measurement ofVAT/SAT, LDL, statin use, PCSK9 genotype, and othersecondary parameters. The intra-observer reproducibilitywas high (intraclass correlation coefficient = 0.99) forboth VAT and SAT measurements. The median value ofthe VAT/SAT ratio was 0.80 (IQR 0.50–1.03). Therewere 36 patients with a calculated VAT/SAT ratio lowerthan 0.80 and these were classified into the low VAT/SAT group, while 37 patients had a calculated VAT/SATratio above 0.80 and were classified into the high VAT/SAT group. The proportion of patients using statins was22.2% (8/36) in the low VAT/SAT group and was 40.5%(15/37) in the high VAT/SAT groups. PCSK9 LOF geno-type was present in 57.1% (20/35) of patients from thelow VAT/SAT group and in 44.1% (15/34) of patientsfrom the high VAT/SAT group.Baseline characteristicsThe baseline characteristics of patients in both lowand high VAT/SAT groups were similar except forage, gender, creatinine, and acute physiology andchronic health evaluation II (APACHE II) scores(Table 1). Compared to low VAT/SAT, patients withhigh VAT/SAT were older (p = 0.025), more oftenmale (p < 0.001), had higher APACHE II scores (p =0.028), and higher creatinine levels at admission (p =0.039). Type of statins and intensity regimen re de-scribed in Additional file 1: Table S1.Relationship between VAT/SAT and LDL levelsSimilar to a previous report in a septic shock cohort [4],patients with a low VAT/SAT ratio (below the median of0.802) had higher 90-day survival in the current sepsiscohort (p = 0.039) (Fig. 1). Patients with a low VAT/SATratio had higher LDL compared to patients having a highVAT/SAT (p = 0.024) (Fig. 2).Effect of statins on LDL levels according to VAT/SATgroupThe beneficial effect of low VAT/SAT was most pro-nounced in control patients not treated with statins(Fig. 3a). Patients with low VAT/SAT not on statintherapy had LDL levels that were significantly higherthan those with high VAT/SAT not on statin therapy(p = 0.006) (Fig. 3a). Thus, low VAT/SAT protectedagainst the very low LDL levels observed in patientswith sepsis. In these patients the higher LDL levels inthe low VAT/SAT group were associated with signifi-cantly higher 90-day survival (p = 0.019) (Fig. 3c).Statin treatment abrogated these beneficial effects. Sta-tin treatment reduced LDL in patients with low VAT/SAT so that patients with low or high VAT/SAT hadsimilarly low LDL levels (p = 0.548) (Fig. 3b). Loss of thedifference in LDL levels in patients with low versus highVAT/SAT was also associated with loss of the differencein 90-day survival (p = 0.774) (Fig. 3d). Thus, statintreatment reveals that decreased LDL production leadsto loss of the LDL-preserving and survival benefit of lowVAT/SAT versus high VAT/SAT.Effect of PCSK9 loss-of-function genotypeSimilar to patients not taking statins, patients withPCSK9 wildtype genotype in the low VAT/SAT grouphad preserved LDL levels that were significantly higherthan those in patients with high VAT/SAT (p < 0.001)(Fig. 4a). Similar to statin treatment, PCSK9 loss-of-function genotype in the patients with low VAT/SATabrogated the effect of VAT/SAT on LDL such that pa-tients with low or high VAT/SAT had similarly low LDLlevels (p = 0.959) (Fig. 4b).In distinct contrast to statin-treated patients, loss ofthe difference in LDL levels by PCSK9 genotype in pa-tients with low versus high VAT/SAT was not associatedwith loss of the survival benefit. In fact, PCSK9 loss-of-function genotype was associated with greater differencebetween improved survival in patients with low VAT/SAT compared to all patients with high VAT/SAT (p =0.017 and p = 0.527) (Fig. 4d versus Fig. 4c). Thus,PCSK9 loss-of-function genotype demonstrates that lowLDL per se does not lead to increased mortality becauselow LDL levels in the PCSK9 loss-of-function genotypepatients with low VAT/SAT was associated with remark-ably high survival. These analyses were performed ac-cording to cholesterol levels at sepsis admission anddemonstrated similar results in comparison to LDL ana-lysis (Additional file 1: Figures S1 to S3).Effect of statins and PCSK9 loss-of-function genotype onLDL levels in the low VAT/SAT groupIn order to rule out a possible beneficial effect ofstatin use on survival, and to confirm ourLee et al. Critical Care  (2018) 22:58 Page 3 of 9hypothesis that LDL clearance (and not decreasedsynthesis) in sepsis confers a survival benefit, weperformed a subgroup analysis only in patients withlow VAT/SAT ratio. In these patients, both statin-treated and PCSK9 LOF genotype groups had sig-nificantly lower LDL compared to the control andwild type (WT) groups, respectively (p = 0.046 andp = 0.028) (Additional file 1: Figures S4A and S4B).However, despite this observation, only the PCSK9LOF group yielded a significant mortality benefitcompared to WT, whereas no appreciable differ-ences in survival were observed between statintreated and untreated control groups (Additional file1: Figure S5A and S5B).Table 1 Characteristics of patients at baselineVariable Visceral to subcutaneous adipose tissue ratio (VAT/SAT)< median (n = 36) ≥ median (n = 37) p valueBaseline characteristicsAge, (IQR) 38–70 56–68 0.025Male, n (%) 18 (50.0) 35 (94.6) <0.001Caucasians, n (%) 31 (93.9)a 20 (80.0) 0.145Chronic condition, n (%)Hypertension 14 (38.9) 12 (32.4) 0.740Congestive heart failure 4 (11.1) 3 (8.1) 0.711Myocardial infarction 3 (8.3) 3 (8.1) 0.136Diabetes mellitus 9 (25.0) 10 (27.0) 1.000COPD 5 (13.9) 11 (29.7) 0.176Chronic kidney disease 2 (5.6) 6 (16.2) 0.261Cirrhosis 1 (2.8) 4 (10.8) 0.358Malignancy 2 (5.6) 2 (5.4) 1.000Hemodynamic variables at admission (IQR)Mean arterial pressure 66–89 72–84 0.911Heart rate 83–104 79–115 0.704Respiratory rateb 16–20 17–22 0.323Temperature (°C) 36.4–37.2 36.3–37.1 0.661Laboratory measurements at admission (IQR)Hemoglobin (g/L) 88.0–130.0 87.0–127.0 0.467WBC (×103/mm3) 5.90–12.97 6.67–17.65 0.220Platelets (109/L) 128–362 119–230 0.075Creatinine (mmol/L) 64–119 77–180 0.039Lactate (mmol/L) 1.1–1.9 1.3–3.3 0.179Other clinical characteristicsAPACHE II score (IQR) 3–12 8–17 0.028Use of vasopressor, n (%) 6 (16.7) 6 (16.2) 0.871Mechanical ventilation, n (%) 12 (33.3) 14 (37.8) 0.444PCSK9 genotype, n (%)One or more loss-of-function 20 (55.6) 15 (40.5) 0.294Two or more loss-of-function 5 (13.9) 9 (24.3) 0.404Adipose tissue area (cm2) (IQR)Visceral adipose tissue 44.7–138.5 126.8–279.5 <0.001Subcutaneous adipose tissue 116.5–270.8 99.8–212.2 0.300Abbreviations: VAT visceral abdominal tissue, SAT subcutaneous adipose tissue, IQR interquartile range, COPD chronic obstructive pulmonary disease, WBC whiteblood cell, APACHE II acute physiology and chronic health evaluation IIaMissing data, n = 3bBreaths per minuteLee et al. Critical Care  (2018) 22:58 Page 4 of 9Effect of statins on survival in the high VAT/SAT groupIn patients with high VAT/SAT, statin treatment did notdemonstrate a survival benefit compared to untreatedcontrol groups (Additional file 1: Figure S6).DiscussionBoth statin treatment and PCSK9 loss-of-function tookaway the LDL-preserving effect of low VAT/SAT resultingin uniformly low LDL across both low and high VAT/SATgroups. Despite this similarity in LDL levels, the survivalrates were not alike. When LDL was lowered using statinsthe survival benefit conferred from having a low VAT/SAT was eliminated. In contrast, PCSK9 loss-of-functiongenotype-driven changes in LDL levels (by increasing LDLclearance) were associated with statistically significantlyimproved outcomes. In other words, while PCSK9 loss-of-function led to the low VAT/SAT group losing their abilityto preserve LDL levels, it did not take away the protectiveeffect, and in fact, enhanced the survival benefit. Thus, itis not the static low LDL levels in sepsis that drive in-creased mortality. Rather, it may be the dynamic interplaybetween LDL production and clearance that contributesto clinical outcomes in sepsis.These VAT/SAT observations in an independent co-hort replicate the finding by Pisitsak et al. that low VAT/SAT is associated with longer survival in sepsis com-pared to high VAT/SAT [4]. Patients with sepsis in thecurrent study were identified earlier in sepsis in theEmergency Department rather than in the ICU andtherefore had less severe illness and lower mortalitycompared to the patients with septic shock who were re-cruited from an ICU, as described in the previous report[4]. In addition to this, we also found that the low VAT/SAT group was able to preserve their LDL levels at base-line (considering no statin treatment and PCSK9 wild-type genotype patients). This preservation of LDL levelswas associated with increased survival. This observationis also concordant with previous studies suggesting thatlow LDL is associated with worse outcomes [7]. How-ever, these previous studies did not examine the effect ofthe underlying determinants of LDL levels – LDL pro-duction (as altered by statins) and LDL clearance (as al-tered according to PCSK9 genotype). Statin therapylowers LDL in the circulation primarily via the inhibitionof LDL production in hepatocytes and, to a lesser extent,via a resulting feedback increase in LDL uptake by up-regulation of LDL receptor expression (12). That statinsprimarily inhibit LDL production is confirmed by theobservation that LDL levels are decreased by statins inpatients with homozygous familial hypercholesterolemia,where LDL receptors are not functioning [13, 14].LDL serves an important role in buffering the responseto endotoxins by forming an endotoxin-lipoprotein com-plex [22, 23]. In vitro studies have demonstrated that lipo-polysaccharide (LPS) binds readily to serum lipoproteins,and that LPS-lipoprotein complexes are less toxic thanunbound LPS [24, 25]. This LPS-lipoprotein-complex for-mation offers protection in an acute setting, but may alsobe implicated in the pathogenesis of atherosclerosis [26].In either scenario, there is value to increasing the clear-ance of toxin-bound lipoproteins because they may other-wise increase the inflammatory response, leading tounfavorable clinical outcomes in sepsis and/or contributeto atherogenesis. In major inflammatory states, there is areduction in plasma levels of LDL in parallel with otherFig. 1 Ninety-day mortality in the low and high visceral adiposetissue/subcutaneous adipose tissue (VAT/SAT) groups. Patients with aVAT/SAT ratio lower than the median value (n = 36) had longersurvival compared to patients with a VAT/SAT ratio higher than themedian value (n = 37) (p = 0.039). This observation is consistent witha previous report in patients with septic shock [4]Fig. 2 Patients with a visceral adipose tissue/subcutaneous adiposetissue (VAT/SAT) ratio lower than the median value (n = 36) hadhigher low-density lipoprotein (LDL) levels (median and interquartilerange, mg/dL) compared to patients with a VAT/SAT ratio greaterthan the median value (n = 37) (p = 0.024)Lee et al. Critical Care  (2018) 22:58 Page 5 of 9lipoproteins [27]. The mechanism behind our observationof patients with low VAT/SAT having the ability to pre-serve their LDL levels (no statin treatment and PCSK9wildtype genotype) compared to those with high VAT/SAT is yet to be fully understood.Although static LDL levels have some value for esti-mating the buffering capacity of the patient against en-dotoxins in sepsis, it does not fully explain the complexdynamic role of LDL production and clearance, whichinteract to determine plasma LDL levels. Our resultsshow that there are marked differences between modu-lating LDL production and LDL clearance in under-standing the association between LDL levels and survivalin sepsis. Thus, it may be incomplete and in some casesincorrect to conclude that absolute LDL levels are of pri-mary importance in sepsis. Rather, decreased productionof LDL may contribute to adverse outcomes but in-creased clearance appears to be associated with benefi-cial outcomes. Perhaps this is because inhibiting theproduction of LDL hinders the buffering capacity of thebody against LPS, whereas increased clearance helpseliminate the LDL-endotoxin complex.Our results further highlight the differential contribu-tions from VAT and SAT in sepsis, which cannot be ex-plained by BMI alone. Several factors that contributedifferently towards VAT and SAT have been established,including gender, age, and the onset of menopause in fe-males [28]. When looking at VAT/SAT ratios, one studysuggested that VAT/SAT ratios are closely associated withage [5], while another found the opposite when looking atyoung Japanese male subjects [29]. As such, our cohortshowed great age variability on analysis according to lowand high VAT/SAT ratios. In parallel with its inflamma-tory properties, adipose tissue is implicated in the devel-opment of metabolic conditions, so adipose tissue is bothan active immune and endocrine organ [30]. VAT is aknown predictor of diabetes mellitus, while SAT is not[31]. An analysis of the Framingham Heart Study foundthat high VAT/SAT ratio was associated with increasedcardiometabolic risk [5]. Our results identify a novel dif-ference between VAT and SAT by modulating lipoproteinlevels differentially during sepsis. Furthermore, both sta-tins [32] and PCSK9 [33] may also be differentially relatedto visceral versus subcutaneous obesity. In health, LDLand other lipoprotein particles are correlated morestrongly with visceral adipose tissue (VAT) than with sub-cutaneous adipose tissue (SAT) [34].Statin use may not improve or worsen survival out-comes on a uniform basis in sepsis and our findings mayhelp elucidate the discrepancy found in currently availableFig. 3 a Untreated control patients not on statins (n = 50) with low visceral adipose tissue/subcutaneous adipose tissue (VAT/SAT) (n = 28) havehigher low-density lipoprotein (LDL) (median and interquartile range, mg/dL) than patients with high VAT/SAT (n = 22) (p = 0.006). b Statin-treatedpatients (n = 23) have low LDL regardless of VAT/SAT category (p value not significant). c Control untreated patients not on statins with low VAT/SAT(n = 28) have longer survival compared to patients with high VAT/SAT (n = 22) (p = 0.019). d In statin-treated patients (n = 23) there is no difference insurvival between VAT/SAT categories (p value not significant)Lee et al. Critical Care  (2018) 22:58 Page 6 of 9statin versus sepsis survival studies [35, 36]. A meta-analysis by Wan et al. showed that prospective studiesfound no benefit of statin use in sepsis survival, whileretrospective studies have demonstrated statin benefit[36]. One prospective double-blind, randomized, con-trolled trial in 2012 (ASEPSIS trial) found that acute ad-ministration of atorvastatin was associated with lessprogression to severe sepsis, but did not change survival[37]. In addition, Beed et al. showed that prior statin usehad no benefit in patients who developed sepsis [38],while other studies report benefit of statins on sepsis sur-vival [39, 40]. Our results demonstrate that VAT/SAT sta-tus – and other factors – may influence the effect ofstatins on survival in sepsis.Limitations of our study include but are not limited tothe inherent biases of a retrospective study design and thatthe design is an association study so causal inference is notpossible. Because our study started before publication ofthe new Sepsis-3 [16] definition we used the previous defin-ition of sepsis in patients at first contact in the EmergencyDepartment. This means that our patients had a wide rangeof severity of illness and included a significant contributionfrom patients without organ failure who had low mortalityrates. A weakness of this is that our patient population ismore heterogeneous than if the Sepsis-3 definition wereused. A strength of this is that our findings may be morebroadly applicable. Patients in our study only had abdom-inal CT scans when the treating physician ordered them forclinical purposes, and this may present an unknown bias.We used PCSK9 loss-of-function genotype as a surrogateto mark PCSK9 inhibition in vivo; however, it is unclearwhether our PCSK9 genotype findings will translate to ef-fects on sepsis outcomes by modulating PCSK9 withPCSK9-inhibiting drugs. We took into account the produc-tion and clearance of LDL in lowering LDL, but there maybe other important mechanisms involved in LDL homeo-stasis (e.g. sequestration by macrophages). Further studieslooking into this these VAT/SAT subgroups and their im-munogenic response, or lack thereof, to statin therapy andthe implications of PCSK9 loss-of-function genotype willbe necessary to elucidate causal pathways.In conclusion, low VAT/SAT is associated with relativepreservation of LDL levels during sepsis and improvedsurvival. However, low LDL levels per se do not appearto cause decreased sepsis survival because inhibitingLDL production with statins (which lowers LDL levels)is associated with adverse outcomes, while loweringplasma LDL by increasing LDL clearance (PCSK9 loss-of-function genotype) was associated with higher sur-vival in patients with a low VAT/SAT.Fig. 4 a Patients who have the wild type genotype (n = 34) with low visceral adipose tissue/subcutaneous adipose tissue (VAT/SAT)(n = 15) have higher low-density lipoprotein (LDL) (median and interquartile range, mg/dL) than patients with high VAT/SAT (n = 19)(p = 0.000082). b Patients who have PCSK9 loss-of-function genotype (n = 35) have low LDL levels regardless of VAT/SAT category (pvalue not significant). c In patients who have wild type genotype (n = 34) there is no significant difference in survival between thosewith low VAT/SAT (n = 15) or high VAT/SAT (n = 19) (p value not significant). d Patients who have PCSK9 loss-of-function genotype(n = 35) with low VAT/SAT (n = 20) have longer survival compared to patients with high VAT/SAT (n = 15) (p = 0.017). Four patientswho carried PCSK9 gain-of-function mutations were excluded from this analysis (n = 69)Lee et al. Critical Care  (2018) 22:58 Page 7 of 9ConclusionsLow LDL levels per se are not simply associated with de-creased sepsis survival, because lowering LDL levels byinhibiting LDL production (statin treatment) is associ-ated with adverse outcomes, while increased LDL clear-ance (PCSK9 loss-of-function genotype) is associatedwith improved outcomes in patients with low VAT/SAT.Increased LDL clearance during sepsis may be a usefultherapeutic goal.Additional fileAdditional file 1: Table S1. Statin subtypes and dosages. Figure S1.Cholesterol levels according to VAT/SAT groups. Higher cholesterol levelswere observed in the low VAT/SAT group (n = 36) than high VAT/SATgroup (n = 37) (p = 0.005). Figure S2A. Cholesterol levels in untreatedcontrol patients according to VAT/SAT groups. In control untreatedpatients (n = 50), low VAT/SAT group (n = 28) had higher cholesterollevels than high VAT/SAT group (n = 22) (p = 0.009). Figure S2B.Cholesterol levels in statin treated patients according to VAT/SAT groups.Patients on statin treatment (n = 23) showed no difference in cholesterollevels between low (n = 8) and high VAT/SAT (n = 15) groups (p valuenot significant). Figure S3A. Cholesterol levels in WT genotype patientsaccording to VAT/SAT groups. Patients with WT genotype (n = 34) hadhigher cholesterol levels in the low VAT/SAT group (n = 15) compared tothe high VAT/SAT group (n = 19) (p = 0.001). Figure S3B. Cholesterollevels in patients with the PCSK9 LOF genotype according to VAT/SATgroup. No differences in cholesterol levels were observed between thelow (n = 20) and high VAT/SAT (n = 15) group in patients with PCSK9LOF (n = 35) (p value not significant). Figure S4A. Patients with low VAT/SAT (n = 36) who have been treated with statins (n = 8) have lower LDLlevels than untreated control patients (n = 28) (p = 0.046). Figure S4B.Patients with low VAT/SAT with PCSK9 LOF genotype (n = 20) have lowerLDL levels than WT patients (n = 15) (p = 0.028). Figure S5A. In patientswith low VAT/SAT, there is no significant difference in survival betweenthe untreated control (n = 28) and statin-treated groups (n = 8) (p =0.485). Figure S5B. Patients with low VAT/SAT who have the PCSK9 LOFgenotype (n = 20) have increased survival compared to patients with WT(n = 15) (p = 0.043). Figure S6. In patients with high VAT/SAT (n = 37),statin treatment did not demonstrate a survival benefit compared to theuntreated control groups (p = 0.410). (DOCX 641 kb)AcknowledgementsKRG is sponsored by CNPq-Brazil.FundingThis study was supported by the Canadian Institutes of Health Research(CIHR).Availability of data and materialsPlease contact the corresponding author for data requests.Authors’ contributionsJGHL and CP made all VAT SAT measurements. All authors contributed tothe design, analysis, and writing and review of the manuscript. All authorsread and approved the final manuscript.Authors’ informationAdditional author information is available at approval and consent to participateAll patients provided written informed consent. This protocol was reviewedand approved by the St. Paul’s Hospital/University of British Columbiaresearch ethics board.Consent for publicationNot applicable.Competing interestsJHB, JAR, and KRW report patents owned by the University of BritishColumbia (UBC) that are related to PCSK9 inhibitor(s) and sepsis. JHB, JAR,and KRW are founders and shareholders of Cyon Therapeutics, a companythat holds patents for use of PCSK9 inhibitors to treat sepsis.JAR reports patents owned by the University of British Columbia (UBC) thatare related to the use of vasopressin in septic shock. Dr. Russell is aninventor on these patents. JAR has share options in Leading Biosciences Inc.JAR is a shareholder in Molecular You Corp. JAR reports receiving consultingfees from:1. Cubist Pharmaceuticals (now owned by Merck; formerly was TriusPharmaceuticals; developing antibiotics),2. Leading Biosciences (developing a sepsis therapeutic),3. Ferring Pharmaceuticals (manufactures vasopressin and is developingselepressin),4. Grifols (sells albumin),5. La Jolla Pharmaceuticals (developing angiotensin II; Dr. Russell chairs theData Safety Monitoring Board (DSMB) of a trial of angiotensin II),6. CytoVale Inc. (developing a sepsis diagnostic),7. Asahi Kesai Pharmaceuticals of America (AKPA) (developing recombinantthrombomodulin).JAR reports having received an investigator-initiated grant from Grifols that isprovided and administered by UBC.No other authors have any conflict of interests.Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1Centre for Heart Lung Innovation, University of British Columbia, Vancouver,BC, Canada. 2Ramathibodi Hospital, Faculty of Medicine, Mahidol University,Bangkok, Thailand.Received: 21 August 2017 Accepted: 8 February 2018References1. Pepper DJ, Sun J, Welsh J, Cui X, Suffredini AF, Eichacker PQ. Increased bodymass index and adjusted mortality in ICU patients with sepsis or septicshock: a systematic review and meta-analysis. Crit Care. 2016;20(1):181. PubMed PMID: 27306751;PubMed Central PMCID: PMCPMC4908772.2. Trivedi V, Bavishi C, Jean R. Impact of obesity on sepsis mortality: a systematicreview. J Crit Care. 2015;30(3):518–524. PMID: 25575851.3. Kuperman EF, Showalter JW, Lehman EB, Leib AE, Kraschnewski JL. Theimpact of obesity on sepsis mortality: a retrospective review. BMC Infect Dis.2013;13:377. PubMed PMID:23957291; PubMed Central PMCID: PMCPMC3765286.4. Pisitsak C, Lee JG, Boyd JH, Coxson HO, Russell JA, Walley KR. Increased ratioof visceral to subcutaneous adipose tissue in septic patients is associatedwith adverse outcome. Crit Care Med. 2016;11:1966–1973. Epub 2016/08/12. PubMed PMID: 27513541.5. Kaess BM, Pedley A, Massaro JM, Murabito J, Hoffmann U, Fox CS. The ratioof visceral to subcutaneous fat, a metric of body fat distribution, is a uniquecorrelate of cardiometabolic risk. Diabetologia. 2012;55(10):2622–2630. Epub2012/08/18. PubMed PMID:22898763; PubMed Central PMCID: PMCPMC3636065.6. Lemieux S, Prud'homme D, Moorjani S, Tremblay A, Bouchard C, Lupien PJ, et al.Do elevated levels of abdominal visceral adipose tissue contribute to age-relateddifferences in plasma lipoprotein concentrations in men? Atherosclerosis 1995;118(1):155–164. Epub 1995/11/01. PubMed PMID: 8579625.7. Shor R, Wainstein J, Oz D, Boaz M, Matas Z, Fux A, et al. Low serum LDLcholesterol levels and the risk of fever, sepsis, and malignancy. Ann Clin LabSci 2007;37(4):343–348. Epub 2007/11/15. PubMed PMID: 18000291.8. Levels JH, Abraham PR, van Barreveld EP, Meijers JC, van Deventer SJ.Distribution and kinetics of lipoprotein-bound lipoteichoic acid. InfectImmun 2003;71(6):3280–3284. Epub 2003/05/23. PubMed PMID: 12761109;PubMed Central PMCID: PMC155762.Lee et al. Critical Care  (2018) 22:58 Page 8 of 99. Levels JH, Abraham PR, van den Ende A, van Deventer SJ. Distribution andkinetics of lipoprotein-bound endotoxin. Infect Immun 2001;69(5):2821–2828. Epub 2001/04/09. PMID: 11292694; PubMed Central PMCID: PMC98230.10. Walley KR, Thain KR, Russell JA, Reilly MP, Meyer NJ, Ferguson JF, et al.PCSK9 is a critical regulator of the innate immune response and septicshock outcome. Sci Transl Med. 2014;6(258):258ra143. Epub 2014/10/17. PubMed PMID: 25320235.11. Dwivedi DJ, Grin PM, Khan M, Prat A, Zhou J, Fox-Robichaud AE, et al. Differentialexpression of PCSK9 modulates infection, inflammation and coagulation in amurine model of sepsis. Shock. 2016;46:672–80. Epub 2016/07/13. PubMed PMID: 27405064.12. Stancu C, Sima A. Statins: mechanism of action and effects. J Cell Mol Med2001;5(4):378–387. Epub 2002/06/18. PubMed PMID: 12067471.13. Marais AD, Naoumova RP, Firth JC, Penny C, Neuwirth CK, Thompson GR.Decreased production of low density lipoprotein by atorvastatin afterapheresis in homozygous familial hypercholesterolemia. J Lipid Res 1997;38(10):2071–2078. Epub 1997/11/28. PubMed PMID: 9374129.14. Raal FJ, Pilcher GJ, Illingworth DR, Pappu AS, Stein EA, Laskarzewski P,et al. Expanded-dose simvastatin is effective in homozygous familialhypercholesterolaemia. Atherosclerosis 1997;135(2):249–256. Epub 1998/01/16. PubMed PMID: 9430375.15. Lepor NE, Kereiakes DJ. The PCSK9 inhibitors: a novel therapeutic targetenters clinical practice. Am Health Drug Benefits 2015;8(9):483–489. PubMedPMID: 26834934; PubMed Central PMCID: PMC4719137.16. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M,et al. The third international consensus definitions for sepsis and septic shock(Sepsis-3). JAMA. 2016;315(8):801–810. PMID: 26903338; PubMed Central PMCID: PMC4968574.17. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, et al.2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis DefinitionsConference. Intensive Care Med 2003;29(4):530–538. Epub 2003/03/29. PubMed PMID: 12664219.18. Stone NJ, Robinson JG, Lichtenstein AH, Bairey Merz CN, Blum CB, Eckel RH,et al. 2013 ACC/AHA guideline on the treatment of blood cholesterol toreduce atherosclerotic cardiovascular risk in adults: a report of the AmericanCollege of Cardiology/American Heart Association Task Force on PracticeGuidelines. J Am Coll Cardiol. 2014;63(25 Pt B):2889-2934. Epub 2013/11/19.doi: PubMed PMID: 24239923.19. Yoshizumi T, Nakamura T, Yamane M, Islam AH, Menju M, Yamasaki K, et al.Abdominal fat: standardized technique for measurement at CT. Radiology1999;211(1):283–286. Epub 1999/04/06. PubMed PMID: 10189485.20. Kvist H, Chowdhury B, Sjostrom L, Tylen U, Cederblad A. Adipose tissuevolume determination in males by computed tomography and 40K. Int JObes 1988;12(3):249–266. Epub 1988/01/01. PubMed PMID: 3391740.21. Rossner S, Bo WJ, Hiltbrandt E, Hinson W, Karstaedt N, Santago P, et al.Adipose tissue determinations in cadavers–a comparison between cross-sectional planimetry and computed tomography. Int J Obes 1990;14(10):893–902. Epub 1990/10/01. PubMed PMID: 2269582.22. Flegel WA, Baumstark MW, Weinstock C, Berg A, Northoff H. Prevention ofendotoxin-induced monokine release by human low- and high-densitylipoproteins and by apolipoprotein A-I. Infection & Immunity. 1993;61(12):5140–6.23. Netea MG, Demacker PN, Kullberg BJ, Boerman OC, Verschueren I, StalenhoefAF, et al. Low-density lipoprotein receptor-deficient mice are protected againstlethal endotoxemia and severe gram-negative infections. J Clin Invest 1996;97(6):1366–1372. Epub 1996/03/15. PubMedPMID: 8617867; PubMed Central PMCID: PMC507194.24. Emancipator K, Csako G, Elin RJ. In vitro inactivation of bacterial endotoxinby human lipoproteins and apolipoproteins. Infect Immun 1992;60(2):596–601. PubMed PMID: 1730494; PubMed Central PMCID: PMC257670.25. Flegel WA, Wölpl A, Männel DN, Northoff H. Inhibition of endotoxin-inducedactivation of human monocytes by human lipoproteins. Infect Immun 1989;57(7):2237–2245. PubMed PMID: 2731990; PubMed Central PMCID: PMC313866.26. Schwartz YSh, Dushkin MI. Endotoxin-lipoprotein complex formation as afactor in atherogenesis: associations with hyperlipidemia and with lecithin:cholesterol acyltransferase activity. Biochemistry (Mosc) 2002;67(7):747–752.PubMed PMID: 12139471.27. Mooser V, Berger MM, Tappy L, Cayeux C, Marcovina SM, Darioli R, et al.Major reduction in plasma Lp(a) levels during sepsis and burns. ArteriosclerThromb Vasc Biol 2000;20(4):1137–1142. PubMed PMID: 10764684.28. Wajchenberg BL. Subcutaneous and visceral adipose tissue: their relation tothe metabolic syndrome. Endocr Rev 2000;21(6):697–738. Epub 2001/01/02. PubMed PMID: 11133069.29. Asayama K, Dobashi K, Hayashibe H, Kodera K, Uchida N, Nakane T, et al.Threshold values of visceral fat measures and their anthropometricalternatives for metabolic derangement in Japanese obese boys. Int JObes Relat Metab Disord 2002;26(2):208–213. Epub 2002/02/19. PubMed PMID: 11850752.30. Bays HE, Gonzalez-Campoy JM, Bray GA, Kitabchi AE, Bergman DA, SchorrAB, et al. Pathogenic potential of adipose tissue and metabolicconsequences of adipocyte hypertrophy and increased visceral adiposity.Expert Rev Cardiovasc Ther 2008;6(3):343–368. Epub 2008/03/11. PubMed PMID: 18327995.31. Bray GA, Jablonski KA, Fujimoto WY, Barrett-Connor E, Haffner S, Hanson RL,et al. Relation of central adiposity and body mass index to thedevelopment of diabetes in the Diabetes Prevention Program. 2008.32. Wee CC, Girotra S, Weinstein AR, Mittleman MA, Mukamal KJ. Therelationship between obesity and atherosclerotic progression and prognosisamong patients with coronary artery bypass grafts the effect of aggressivestatin therapy. J Am Coll Cardiol. 2008;52(8):620–5.33. Arsenault BJ, Pelletier-Beaumont E, Almeras N, Tremblay A, Poirier P,Bergeron J, et al. PCSK9 levels in abdominally obese men: association withcardiometabolic risk profile and effects of a one-year lifestyle modificationprogram. Atherosclerosis 2014;236(2):321–326. Epub 2014/08/17. PubMed PMID: 25128757.34. Deschenes D, Couture P, Dupont P, Tchernof A. Subdivision of thesubcutaneous adipose tissue compartment and lipid-lipoprotein levels inwomen. Obes Res 2003;11(3):469–476. Epub 2003/03/14. PubMed PMID: 12634447.35. Kopterides P, Falagas ME. Statins for sepsis: a critical and updated review.Clin Microbiol Infect 2009;15(4):325–334. Epub 2009/05/07. PubMed PMID: 19416304.36. Wan YD, Sun TW, Kan QC, Guan FX, Zhang SG. Effect of statin therapy onmortality from infection and sepsis: a meta-analysis of randomized andobservational studies. Crit Care. 2014;182014:R71.37. Patel JM, Snaith C, Thickett DR, Linhartova L, Melody T, Hawkey P, et al.Randomized double-blind placebo-controlled trial of 40 mg/day of atorvastatinin reducing the severity of sepsis in ward patients (ASEPSIS Trial). Crit Care.2012;16(6):R231. Epub 2012/12/13. PubMedPMID: 23232151; PubMed Central PMCID: PMCPMC3672620.38. Beed M, Brindley PG, Mahajan R, Juttner I, Campion-Smith J, Wilson VG.The association between prior statin use and long-term outcomes aftercritical care admission. J Crit Care 2016;35:63–68. Epub 2016/08/03. PubMed PMID: 27481737.39. Almog Y, Shefer A, Novack V, Maimon N, Barski L. Eizinger M, et al. Priorstatin therapy is associated with a decreased rate of severe sepsis. 2004; Fuller BM, Gajera M, Schorr C, Gerber D, Dellinger RP, Zanotti S. Theassociation of prior statin use in septic shock treated with early goaldirected therapy. Eur J Emerg Med. 2012;19(4):226–230. Epub 2011/09/15. PubMed PMID: 21915053;PubMed Central PMCID: PMCPMC3556478.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript your next manuscript to BioMed Central and we will help you at every step:Lee et al. Critical Care  (2018) 22:58 Page 9 of 9


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items