Cholesterol Rather Than Procalcitonin or C-Reactive Protein Predicts Mortality in Patients With Infection
Biller, Katharina*; Fae, Peter†; Germann, Reinhard†; Drexel, Heinz‡; Walli, Autar K.§; Fraunberger, Peter*
*Medical Central Laboratories; †Clinic of Anesthesiology and Intensive Care Medicine, Academic Teaching Hospital; and ‡Vorarlberg Institute for Vascular Investigation and Treatment (VIVIT), Feldkirch, Austria; and §Department of Clinical Chemistry, University Hospital Großhadern, Munich, Germany
Received 3 Feb 2014; first review completed 17 Feb 2014; accepted in final form 2 Apr 2014
Address reprint requests to Peter Fraunberger, MD, Medical Central Laboratories, Carinagasse 41, A-6800 Feldkirch, Austria. E-mail: firstname.lastname@example.org.
ABSTRACT: Serum cholesterol procalcitonin (PCT) and C-reactive protein (CRP) levels were measured consecutively in 76 critically ill patients at admission to the intensive care unit. The presence of infection was defined according to the CDC (Centers for Disease Control and Prevention) criteria; in-house mortality, underlying diseases, and severity of sepsis were monitored. Nonsurvivors had significantly lower cholesterol levels compared with survivors (69 mg/dL [range, 37–88 mg/dL] vs. 96 mg/dL [range, 71–132 mg/dL], P = 0.006) whereas no significant differences were noted for serum PCT and CRP levels. In a cohort of patients with cholesterol levels of 50 mg/dL or less, 82% did not survive as compared with patients with cholesterol levels of 100 mg/dL or greater (mortality, 21%). In a control group without infection, no difference of cholesterol, PCT, or CRP was found between survivors and nonsurvivors. Our data show that low cholesterol levels in patients with infectious disease have a prognostic value and may be useful markers to identify high-risk patients already at admission.
Infection is one of the major reasons for admission to the intensive care unit and mortality. According to the American College of Chest Physicians and the Society of Critical Care Medicine (ACCP/SCCM), infection is defined as an invasive process by pathogenic or potentially pathogenic microorganisms to normally sterile tissue, fluids, or body cavities (1). As a consequence, the severe systemic inflammation response of the organisms to infection is defined as sepsis. Even though therapeutic measures have considerably improved in recent years, yet mortality of sepsis has not decreased. Delayed recognition with subsequent inadequate therapy is a major reason for this. Therefore, early markers to identify patients at high risk for sepsis are needed.
Various biomarkers have been tested for diagnosis of sepsis, evaluation of therapy success, and prognostic value in intensive care patients (2). In particular, C-reactive protein (CRP) and, during the last decade, procalcitonin (PCT) levels are frequently used in clinical practice because both are elevated during inflammatory processes. C-reactive protein is a marker to detect local infection and has been used for monitoring inflammatory response to therapy. However, for diagnosis of sepsis, sensitivity and specificity are low. Therefore, PCT has been proposed to be a more specific marker for bacteremia and sepsis. Furthermore, because of the high negative predictive value, decrease in PCT levels may be an indicator of successful antibiotic therapy (3). However, various studies indicate that the prognostic value of both CRP and PCT in patients with infection and sepsis is limited because of false-positive or negative values (4).
In addition to markers such as CRP and PCT, which represent the immune response of the organism, metabolic changes during inflammatory disease have also been found to be of prognostic value in patients with infection and sepsis. Already in 1911, Chauffard et al. (5) reported a decrease in serum cholesterol levels in febrile patients. Since then, a number of studies have reported hypocholesterolemia during various bacterial or viral infectious diseases, and various studies indicate a correlation between cholesterol decrease and adverse outcome during inflammatory diseases in hospitalized patients (6). For example, in children with severe meningococcal sepsis, high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C) levels on admission were inversely associated with severity of disease and mortality (7). Likewise, in patients with bacterial infection of the lower respiratory tract, low cholesterol levels were predictive for adverse outcome (8).
Studies in humans, animal models as well as cells in culture show that lipoproteins such as LDL, oxidized LDL, Lp(a), HDL, or apolipoprotein A1 reduce inflammatory response to bacteria and bacterial products (9). Recent data from studies on LDL receptor knockout mice demonstrate that high circulating lipoprotein levels protect against infection induced by lipopolysaccharide (LPS) or Salmonella (10). Whether cholesterol decrease in human sepsis is a secondary manifestation of disease or actively contributes to deterioration during infectious diseases is unclear. In the present study, we attempt to delineate the prognostic role of circulating cholesterol in intensive care patients with infectious diseases. Our data indicate a prognostic predictive value of circulating cholesterol in patients with infections.
MATERIALS AND METHODS
Intensive care unit patients with diagnosis of infection (n = 76) were consecutively included in this prospective study. Patients admitted to the intensive care unit from 2007 until 2011 fulfilling criteria of the ACCP/SCCM consensus conference were included in the study. The presence of infection was defined according to the CDC (Centers for Disease Control and Prevention) criteria including microbiological proof and one of the following criteria: elevated CRP or clinical signs of infection (fever, shivering, local signs) or radiological findings. In addition, in-house mortality, underlying diseases, and severity of sepsis were monitored (1).
A control group of 40 intensive care patients without infections was included for comparison. The study was approved by the local ethics committee.
Blood was drawn immediately after admission to the intensive care unit, centrifuged, aliquoted, and kept frozen at −70°C for analysis. Cholesterol and CRP levels were measured with Roche automated system (Integra 800; Roche diagnostics, Vienna, Austria). Procalcitonin was measured with an immunoluminometric assay (LUMItest, Fa; Brahms Diagnostica, Berlin, Germany).
All plasma levels are given as median and interquartile range. Comparisons between survivors and nonsurvivors regarding cholesterol were based on one-way analysis of variance, due to a Gaussian distribution of the measured data for cholesterol. Comparisons between survivors and nonsurvivors regarding CRP and PCT were based on Mann-Whitney-Wilcoxon test, because the collected data do not follow a normal distribution. The Kolmogorov-Smirnov test was used to asses the distribution of the data. Receiver operating characteristic (ROC) curve analysis was used for identification of optimal cutoff points and to examine discrimination, and areas under the curve (AUCs) were compared (11). For the ROC curve, data for cholesterol were transformed by 1/value to have all the curves in the same direction. P < 0.05 was regarded as statistically significant.
Patient characteristics are summarized in Table 1. Infectious diseases of patients included pneumonia (n = 41), cholecystitis (n = 1), peritonitis (n = 20), urinary tract infection (n = 1), hemolytic uremic syndrome (n = 1), wound infection (n = 2), spondylodiscitis with staphylococcal sepsis (n = 3), pleuraempyema (n = 2), infections associated with medical devices (n = 3), and pancreatitis (n = 2). Among the specimens, tracheal fluid and bronchoalveolar lavage (70%) was the most common, followed by pus/tissue aspirate (20%), catheter tips (6%), blood (4%), and vaginal fluid (2%).
In three patients with infections, no specimens were obtained at all (the diagnosis was made by radiological findings); in further three patients, no pathogens could be verified, 36 patients had bacterial infections, three had fungal infections, and 31 patients had simultaneous fugal and bacterial infections. Overall, gram-positive infections were more frequent than gram-negative infections. Furthermore, aerobic bacterial infections were more frequent than anaerobic bacterial infections. Most frequently verified bacteria were Enterococcus species, E. coli, coagulase-negative Staphylococci, Streptococcus species, Staphylococcus aureus, Klebsiella species, and Haemophilus influenzae. Most frequently verified fungi were Candida albicans, Candida tropicalis, and Candida glabrata. According to the ACCP-SCCM criteria, 24 patients fulfilled criteria of sepsis, 15 patients had severe sepsis, and 37 patients had septic shock.
In the whole cohort 32% (n = 37) of patients did not survive. In the cohort of the infected patients, 38% (n = 29) of patients did not survive. Cholesterol levels of nonsurvivors were significantly lower compared with patients who survived (P = 0.006: 69 mg/dL [range, 37–88 mg/dL] vs. 96 mg/dL [range, 71–132 mg/dL]; Fig. 1 A). Procalcitonin and CRP levels showed no significant differences between nonsurvivors and survivors (PCT: 2.5 ng/mL [range, 0.40–10.68 ng/mL] vs. 2.4 ng/mL [range, 0.60–16.76 ng/mL]; CRP: 86.5 mg/L [range, 38.7–196.7 mg/L] vs. 154.2 mg/L [range, 57.1–271.9 mg/L]; respectively). In the control group without infection, no difference between survivors and nonsurvivors could be found for cholesterol, PCT, and CRP levels (123 mg/dL [range, 80–148 mg/dL] vs. 114 mg/dL [range, 106–127 mg/dL]; 0.17 ng/dL [range, 0.06–0.83 ng/dL] vs. 0.15 ng/dL [range, 0.04–0.70 ng/dL]; 47.0 mg/L [range, 11.3–96.0 mg/L] vs. 43.5 mg/L [range, 2.9–181.0 mg/L], respectively).
In patients with infection, cholesterol level of 80 mg/dL was associated with a sensitivity of 72%, a specificity of 66%, a positive predictive value of 55%, and a negative predictive value of 80%. Receiver operating characteristic analysis revealed AUC of 0.715 for cholesterol and survival. Receiver operating characteristic analysis for CRP and PCT levels revealed AUCs of 0.407 and 0.474, respectively. In a cohort of patients with cholesterol levels of 50 mg/dL or less, 82% of patients did not survive, whereas patients with cholesterol levels of 100 mg/dL or greater showed a mortality of only 21% (Fig. 1B).
Although PCT and CRP levels are generally recognized diagnostic markers for detection of inflammation and infection, they failed to predict mortality rates in our setting. In accord with current literature, CRP levels are predictors of response to therapy rather than prognostic markers (2). The role of PCT as a prognostic marker, however, is still under debate. Some studies show no association of PCT levels and survival, whereas others found an association between an increase in PCT levels during hospital stay and high mortality (12, 13). In particular, during the early phase of inflammation, prognostic potential of PCT levels is doubtful (14).
Early identification of high-risk patients in the intensive care unit is essential for diagnostic and therapeutic management. Therefore, we investigated the possible role of cholesterol levels as a prognostic marker on the first day of admission to the intensive care unit because cholesterol levels have been reported to be of prognostic value irrespective of the underlying disease in a large population of hospitalized patients (15). The underlying mechanisms for the rapid fall of cholesterol are still unclear. Probably increased demand of cholesterol as well as impaired synthesis due to liver dysfunction may account for this decrease (16).
In our study, cholesterol decrease has a prognostic value only in patients with infection but not in patients without infection. These data support the hypothesis that circulating lipoproteins, such as LDL or HDL, which carry more than 90% of circulating cholesterol, may play a protective role during infectious disease. In vitro studies demonstrated that lipoproteins bind to the bioactive lipid a portion of LPS, thereby decreasing the bioavailability of this toxin to various endotoxin responsive cells (9). Low-density lipoprotein receptor–deficient (LDLR−/−) mice, with a sevenfold increase in plasma LDL levels, are resistant against infection with Salmonella typhimurium compared with LDLR+/+ mice (10). Lipoproteins interfere with binding of Salmonella to host cells and thus provide a scavenger mechanism for LPS. Even though cholesterol plays a protective, anti-inflammatory role in intensive care patients, its circulating levels cannot be significantly increased by infusion or dietary means. Therefore, another approach to reduce inflammatory response would be to suppress activation of nuclear factor κB, which is essential for the release of proinflammatory cytokines. A possible new drug for this purpose are HMG-CoA reductase inhibitors (statins). In addition to lowering cholesterol, these drugs have a pleiotropic anti-inflammatory effect due to suppression of farnesylated isoprenoids essential for signal transduction and subsequent activation of nuclear factor κB (17). Because of a possible protective role of circulating lipoproteins in infectious disease, use of statins as an adjuvant therapy in sepsis seems a paradox, because of its cholesterol-lowering effect. However, recently, it has been shown, both in studies on experimental animals and first clinical studies, that treatment with statins may be beneficial in severe inflammation and sepsis despite lowering circulating cholesterol (18–21).
In summary, our data show that cholesterol levels delineate between survivors and nonsurvivors in patients with infectious disease. Therefore, monitoring of circulating cholesterol in patients with infectious disease or sepsis is useful to identify high-risk patients for mortality and may help to optimize clinical approach for treatment of these patients.
1. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, Cohen J, Opal SM, Vincent JL, Ramsay G: SCCM/ESICM/ACCP/ATS/SIS. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med 3: 1250–1256, 2003.
2. Pierrakos C, Vincent JL: Sepsis biomarkers: a review. Crit Care 14: R15, 2010.
3. Schuetz P, Albrich W, Mueller B: Procalcitonin for diagnosis of infection and guide to antibiotic decisions: past, present and future. BMC Med 9: 107, 2011.
4. Biller K, Fae P, Germann R, Walli AK, Fraunberger P: How reliable is procalcitonin as antiinflammatory marker? J Lab Med 35: 343–348, 2011.
5. Chauffard A, Richet C, Grigaut A: La Cholesterineme au cours de la Tuberculose pulmonaire. Compt Rend Soc Boil ixx: 276, 1911.
6. Wendel M, Paul R, Heller AR: Lipoproteins in inflammation and sepsis. II. Clinical aspects. Intensive Care Med 16: 25–35, 2007.
7. Vermont CL, den Brinker M, Kâkeci N, de Kleijn ED, de Rijke YB, Joosten KF, de Groot R, Hazelzet JA: Serum lipids and disease severity in children with severe meningococcal sepsis. Crit Care Med 33: 1610–1615, 2005.
8. Gruber M, Christ-Crain M, Stolz D, Keller U, Müller C, Bingisser R, Tamm M, Mueller B, Schuetz P: Prognostic impact of plasma lipids in patients with lower respiratory tract infections—an observational study. Swiss Med Wkly 139: 166–172, 2009.
9. Berbée JF, Havekes LM, Rensen PC: Apolipoproteins modulate the inflammatory response to lipopolysaccharide. J Endotoxin Res 11: 97–103, 2005.
10. Netea MG, Joosten LA, Keuter M, Wagener F, Stalenhoef AF, van der Meer JW, Kullberg BJ: Circulating lipoproteins are a crucial component of host defence against invasive Salmonella typhimurium
infection. PLoS One 4: e4237, 2009.
11. Metz CE: Basic principles of ROC analysis. Semin Nucl Med 8 (4): 283–298, 1978.
12. Harbarth S, Holeckova K, Froidevaux C, Pittet D, Ricou B, Grau GE, Vadas L, Pugin J Geneva Sepsis Network. Diagnostic value of procalcitonin, interleukin-6, and interleukin-8 in critically ill patients admitted with suspected sepsis. Am J Respir Crit Care Med 164: 396–402, 2001.
13. Ruiz-Alvarez MJ, García-Valdecasas S, de Pablo R, Sanchez García M, Coca C, Groeneveld TW, Roos A, Daha MR, Arribas I: Diagnostic efficacy and prognostic value of serum procalcitonin concentration in patients with suspected sepsis. J Intensive Care Med 24: 63–71, 2009.
14. Kibe S, Adams K, Barlow G: Diagnostic and prognostic biomarkers of sepsis in critical care. J Antimicrob Chemother 66 (Suppl 2): ii33–ii40, 2011.
15. Windler E, Ewers-Grabow U, Thiery J, Walli A, Seidel D, Greten H: The prognostic value of hypocholesterolemia in hospitalized patients. Clin Investig 72: 939–943, 1994.
16. Chiarla C, Giovannini I, Giuliante F, Zadak Z, Vellone M, Ardito F, Clemente G, Murazio M, Nuzzo G: Severe hypocholesterolemia in surgical patients, sepsis, and critical illness. J Crit Care 25: 361, 2010.
17. Jasińska M, Owczarek J, Orszulak-Michalak D: Statins: a new insight into their mechanisms of action and consequent pleiotropic effects. Pharmacol Rep 59: 483, 2007.
18. Fraunberger P, Gröne E, Gröne HJ, Walli AK: Simvastatin reduces endotoxin-induced nuclear factor kappaB activation and mortality in guinea pigs despite lowering circulating low-density lipoprotein cholesterol. Shock 32: 159, 2009.
19. Zhang S, Luo L, Wang Y, Rahman M, Lepsenyi M, Syk I, Jeppsson B, Thorlacius H: Simvastatin protects against T cell immune dysfunction in abdominal sepsis. Shock 38: 524–531, 2012
20. Azevedo LC, Park M, Schettino GP: Novel potential therapies for septic shock. Shock 30 (Suppl 1): 60–66, 2008.
21. Piechota M, Barylski M, Hannam S, Piechota-Urbańska M, Banach M: Rationale of statin therapy in septic patients. Curr Vasc Pharmacol 11: 795, 2013.
Cholesterol; prognosis; infection; mortality; procalcitonin; C-reactive protein
© 2014 by the Shock Society
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.
Readers Of this Article Also Read