Prognostic value of inflammatory markers (notably cytokines and procalcitonin), nutritional assessment, and organ function in patients with sepsis

ARTICLE

Auteur(s) : Marta Suárez-Santamaría¹, Francisco Santolaria², Alina Pérez-Ramírez², María-Remedios Alemán-Valls², Antonio Martínez-Riera², Emilio González-Reimers², María-José de la Vega¹, Antonio Milena¹

¹Biochemistry laboratory
²Department of Internal Medicine, University Hospital of the Canary Islands, Tenerife, Spain

accepté le 14 Octobre 2009

Sepsis is defined as a systemic inflammatory response of the host to infection. Excessive immune response, with overproduction of inflammatory mediators, results in tissue damage and organ dysfunction [1, 2]. Sepsis presents a wide range of severity and is one of the major causes of death. Hence the need for precise staging and prognostic assessment based on clinical data and on easily available laboratory parameters [3-5].

Procalcitonin (PCT) and C-reactive protein (CRP) have proved useful in the diagnosis of bacterial infection and sepsis [6-9]. Like other mediators of the inflammatory response, PCT correlates with the severity of sepsis. PCT may also serve to discriminate between severe sepsis and less severe forms of infection [10, 11]. However, there is some controversy about its prognostic value for mortality compared with other inflammatory markers, organ dysfunction, and with disease severity scores such as Sequential Organ Failure Assessment (SOFA) [12-26].

The objective of this study was to analyse the prognostic value for short-term mortality of serum PCT levels, and of other mediators of the inflammatory response, such as CRP, pro-inflammatory (TNF-α and IL-6) and anti-inflammatory [IL-10 and IL-1 receptor antagonist (IL-1ra)] cytokines, lipopolysaccharide-binding protein (LBP), soluble receptor CD14 (sCD14), and triggering receptor expressed on myeloid cells-1 (TREM1) in patients hospitalized for sepsis. We also compared its prognostic value with those of nutritional status and organ dysfunction.

Patients and methods

Clinical evaluation

Sepsis was defined as clinical evidence of infection, plus two or more of the following criteria for systemic inflammatory response syndrome (SIRS): 1) temperature > 38° or < 36°C; 2) heart rate > 90 beats/m; 3) respiratory rate > 20/m or PaCO₂ < 32 mmHg; and 4) WBC > 12.000 or < 4.000 cells/mm³, or > 10% immature forms). Sepsis was considered severe when it was associated with hypoperfusion, or organ dysfunction such as metabolic acidosis, oliguria, or an acute alteration in mental status (otherwise, the diagnosis was non-complicated sepsis). Septic shock was diagnosed when hypotension (systolic blood pressure of < 90 mmHg or a reduction of ≥ 40 mmHg from baseline) persisted despite adequate fluid therapy or when vasoactive amines were needed [2]. Finally, the sequential organ failure assessment (SOFA) score was determined at inclusion, and organ failure was considered as a score of 10 points or more [27]. Seventy two cases were diagnosed with non-complicated sepsis, 115 severe sepsis, 43 septic shock and 23 with organ failure.

Comorbidity

Sepsis localization

Organ function was assessed by the SOFA score, which is based on six criteria: 1) respiratory function assessed by PaO₂/FiO₂; 2) coagulation by platelet count; 3) liver function by bilirubin levels; 4) cardiovascular function by the presence of hypotension and the need for dopamine, dobutamine or norepinephrine; 5) central nervous system function by the Glasgow coma scale; and 6) renal function by oliguria and serum creatinine levels [27, 28].

Nutritional assessment

Subjective nutritional evaluation (SNS) included examination of the muscle mass of the upper and lower limbs and of the temporal muscle, defining two degrees of atrophy (severe, moderate), and absence of atrophy; we assigned 2, 1 and 0 points to each category. Bichat’s fat and subcutaneous fat atrophy, recorded by physical examination of the fat loss on the cheek and abdomen, respectively, were scored in the same way. Thus, we obtained a subjective nutritional score based on the sum of the assigned points. As previously reported, a score of 0 to 2 was considered normal, one of 3-4 points as mild malnutrition, and higher than 4 points as severe malnutrition [29].

Mediators of the inflammatory response

Statistical analysis (SPSS15.0)

Results

As shown in table 2, the variables most closely related with 28-day mortality were organ dysfunction (SOFA and Glasgow coma scores), serum lactate, ferritin and LDH levels, and nutritional data such as SNS, handgrip strength and serum transferrin levels, whereas inflammatory markers showed only a moderate correlation. Among the inflammatory markers, elevated IL-10 and IL-6 showed a higher correlation with mortality than increased PCT, TNF, IL-1ra and TREM-1 and decreased LBP, whereas CRP and sCD14 were not related to mortality.

The relationship between PCT and mortality was weaker than those of the other inflammatory markers. Regarding the time course, the relationship between PCT and mortality was only significant at day 28, whereas TNF, TREM-1, IL-1ra, IL-6 and IL-10 were also related to mortality at day 7 and at the end of the hospital stay; IL-6 and IL-10 were also related to mortality after six months and one year. LBP, which increased with the severity of sepsis, was related to mortality at day 28, but, paradoxically, patients who died showed lower serum LBP levels.

Procalcitonin levels were significantly related to the severity of sepsis. Patients with a PCT > 0.5 ng/mL had a relative risk [RR (CI 95%)] of 4.1 (2.2-8.1) for either severe sepsis or worse; as PCT increased, the risk increased. For PCT > 3.3 ng/mL, the RR was 24.3 (3.3-180), whereas none of the patients whose PCT reached 10 ng/mL had non-complicated sepsis. Also, IL-6 and IL-10 were closely related to severity; IL-6 >170 pg/mL had a RR of 17.2 (2.3-127) and IL-10 > 25 pg/mL had a RR of developing severe sepsis, septic shock or organ failure of 14.6 (2-109).

Blood cultures were performed in 205 patients, and 36 of them (18%) were positive after excluding contaminants. Positive blood cultures were not directly related to mortality, but to severity of sepsis. Only 3% of patients with non-complicated sepsis showed positive blood cultures, whereas in those with severe sepsis, or a more advanced form, blood cultures were positive in 20.5% of cases. Positive blood cultures were found in 53% of patients with a SOFA score > 10 (p = 0.001). Patients with positive blood cultures showed increased levels of inflammatory markers. The best inflammatory marker for predicting positive blood culture was PCT, with an AUROC of 0.693 (0.598-0.787), followed by TNF-α, 0.677 (0.585-0.768) and IL-1ra, 0.664 (0.572-0.755). Only 6% of patients with a PCT < 0.5 ng/mL had positive blood cultures, whereas 35% of those with a PCT ≥ 10 ng/mL showed positive blood cultures (p = 0.001).

Regarding the origin of the sepsis, the most frequent focus was the lung, with 140 cases (55%). Pulmonary sepsis was more frequent in men (62%) than in women (46%) (p = 0.013). Patients with pulmonary sepsis had more predisposing factors, worse nutritional status and a blunted inflammatory response with lower serum cytokine levels (table 3). Blood cultures were also less frequently positive (11% versus 24%) in pulmonary sepsis than in non-pulmonary sepsis (p = 0.025), whereas organ failure (SOFA) and death at 28 days (21% versus 18%) - or during the hospital stay (24% versus 19%) - were not significantly different. However, pulmonary sepsis was related to severe comorbidities such as dementia, cancer, liver cirrhosis and immunosuppressive therapy (including steroids). So, despite a similar age, pulmonary sepsis was related to higher, long-term mortality after six month or one year. Moreover, regarding the capacity to predict 28^-day mortality in sepsis according to origin, inflammatory markers, nutritional assessment and organ function data had better predictive value in non-pulmonary than in pulmonary sepsis (table 4). The following parameters showed a greater area under the ROC curve (non-pulmonary/pulmonary), PCT (0.679/0.534) IL-10 (0.792/0.626), IL-6 (0.774/0.619) serum transferrin (0.826/0.753), serum ferritin (0.817/0.673), lactate (0.837/0.756) and SOFA (0.854/0.737).

PCT only showed a moderate relationship to mortality. To determine whether it had independent prognostic value for mortality, we performed a logistic regression analysis including PCT, IL-6, IL-10 with SNS and SOFA. Procalcitonin was bettered by IL-10, SNS and SOFA.

We did not find increased serum leptin levels in patients with sepsis as compared with controls. There were no significant differences in these levels between non-complicated sepsis and severe sepsis; however, patients with septic shock and organ failure showed significantly decreased serum leptin levels when compared to patients with less severe forms of sepsis (p < 0.001). Serum leptin levels correlated directly to nutrition (BMI, SNS, serum cholesterol, albumin prealbumin and transferrin), and inversely with serum IL-6 levels. Patients who died showed lower serum leptin levels than those who survived.
Table 1 Inflammatory markers in the diverse stages of sepsis

Discussion

However, there is some controversy about the prognostic value of PCT for mortality. While some authors [6, 19, 20, 22-24, 26, 35, 36] have found some predictive ability of PCT (determined at day 1 of the onset of sepsis), others have not [12, 34]. Yet others have reported that PCT determined at day 1 showed no prognostic value, but PCT determined later did correlate with mortality [13-16]. Dahaba et al. (2006), in a study of postoperative patients with severe sepsis, reported that the best discriminative value of PCT for mortality is reached at day 6 [15]. Regarding the possible independent predictive value of PCT, Phua et al. (2007) reported that IL-6, IL-10, lactate and SOFA were better outcome predictors than PCT determined on the first day of sepsis; Heper et al. (2006) reported that IL-10 and TNF were better predictors of mortality than PCT; Lee et al. (2008) found that the predictive value of PCT is surpassed by Mortality in Emergency Department Score (MEDS); and Ruiz-Alvarez et al. (2009) reported that, on multivariate analysis, PCT was displaced by CRP, and SOFA [16, 21, 25, 37]. However, Meng et al. (2009), reported that a PCT of over 10 ng/mL was a better predictor of outcome than APACHEII [36].

Our results showed inflammatory markers to be moderately useful as predictors of 28^-day mortality (table 2), but the area under the receiver operating characteristics curve was less than 0.700 in all cases, whereas nutritional data (subjective nutritional score), handgrip or serum transferrin), and organ damage data (LDH, lactate, ferritin or SOFA), all showed an area under the receiver operating curve of over 0.700. According to our data and regarding mortality, PCT was not an ideal inflammatory marker, being surpassed by IL-6 and IL-10. Moreover, PCT showed no independent prognostic value, as on multivariate analysis it was displaced by SNS, SOFA and IL-10. The relationship between PCT and mortality was only significant at day 28; it was not significant at other times (day 7, at discharge, or after six months and one year from inclusion). Neither did the prognostic value of PCT improve when we considered PCT values over 10 ng/mL.

Pulmonary sepsis was observed in half of our patients with sepsis, and the differences with cases of non-pulmonary sepsis are noteworthy (table 4). Patients with pulmonary sepsis showed worse nutrition, less frequent bacteremia and a blunted inflammatory response, with a less marked increase in inflammatory mediators. However, there were not many differences regarding organ function (decreased creatinine and BUN were rather related to impaired nutrition and decreased muscle mass). Early death (at day 7 and 28 or during the hospital stay), was similar in pulmonary and non-pulmonary sepsis patients. However, patients with pulmonary sepsis were more frequently affected by some of the severe predisposing factors such as cancer, dementia, cirrhosis, immunosuppression and malnutrition, so, despite similar age, pulmonary sepsis was related to higher, later mortality rates (after six months or one year).

These differences between pulmonary and non-pulmonary sepsis probably explain, in part, why the prognostic factors are more useful in non-pulmonary sepsis. Whereas outcome in pulmonary sepsis may be more related to severe predisposing conditions, such as cancer or dementia, in non-pulmonary sepsis, inflammatory mediators, malnutrition and data for impaired organ function were closely related to the prognosis. Regarding the capacity of inflammatory markers, nutritional assessment and organ dysfunction data to predict outcome after 28 days, the area under the ROC curve was greater in non-pulmonary sepsis. The differences observed between pulmonary and non-pulmonary sepsis indicate two different types of sepsis. One, pulmonary, in which severe predisposing factors are frequent, with impaired nutrition, less bacteremia, a less intense inflammatory response and higher long-term mortality, and another, non-pulmonary, with fewer underlying factors, better nutrition and a greater inflammatory response.

Two of the inflammatory markers, CRP and sCD14, did not show any relationship with severity of sepsis or with 28-day mortality. CRP is a widely used diagnostic marker of sepsis, but has not shown a strong relationship with mortality [17, 26, 35, 38]. CD14 receptor may be detached from the membrane and can be determined as soluble CD14. Landmann et al. (1995) and Burgmann et al. (1996), reported that patients who died from sepsis showed raised serum sCD14 levels [39, 40]. In accordance with other studies on pneumonia, we found no differences regarding survival [41].

Lipopolysaccharide binding protein (LBP) showed paradoxical behaviour. Whereas LBP levels increase in sepsis, especially in patients with bacteremia, and directly correlated with other inflammatory mediators [42], we found that decreased LBP related to poor prognosis: patients who died had lower values than patients who survived, as reported by Opal et al. (1999) [43]. However, these results have not been confirmed by other studies [41, 44].

There is some controversy about leptin in sepsis. Maruna et al. (2001) [45] consider it to be an acute phase reactant, as it increases in post-surgical sepsis. However, in our patients, serum leptin levels were related to nutrition and inversely correlated with IL-6 levels. Patients with septic shock and organ failure, and those who died, showed lower serum leptin levels than those who survived. These results are similar to those reported by Aleman et al. (2002) in lung cancer and Diez (2008) in community-acquired pneumonia patients [46, 47]. So, we suggest that in patients with sepsis, leptin should be considered as a nutritional parameter rather than as an inflammatory marker.

In conclusion, the prognostic value of PCT and other inflammatory markers was only moderate, being surpassed by nutritional and organ function data. The best inflammatory markers related to mortality were IL-6 and IL-10, rather than PCT, which was only poorly related. However, among the inflammatory markers, PCT proved the best predictor of bacteremia and progression of sepsis, showing greater ability at discriminating between non-complicated and more severe forms of sepsis. The markers sCD14 and CRP showed no ability to predict outcome.

Acknowledgments

Disclosure. None of the authors has any conflict of interest to disclose.

References

2 Bone RC. Toward an epidemiology and natural history of SIRS. JAMA 1992; 268: 3452-5.

3 Vincent JL, Mercan D. Dear Sirs, what is your PCT? Intensive Care Med 2000; 26: 1170-1.

4 Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International sepsis definitions conference. Crit Care Med 2003; 31: 1250-6.

5 Castelli GP, Pognani C, Meisner M, Stuani A, Bellomi D, Sgarbi L. Procalcitonin and C-reactive protein during systemic inflammatory response syndrome, sepsis and organ dysfunction. Crit Care 2004; 8: R234-R242.

6 Müller B, Becker KL, Schächinger H, et al. Calcitonin precursors are reliable markers of sepsis in a medical intensive care unit. Crit Care Med 2000; 28: 977-83.

7 Harbarth S, Holeckova K, Froidevaux C, et al. Diagnostic value of procalcitonin, interleukin-6, and interleukin-8 in critically ill patients admitted with suspected sepsis. Am J Respir Crit Care Med 2001; 164: 396-402.

8 Luzzani A, Polati E, Dorizzi R, Rungatscher A, Pavan R, Merlini A. Comparison of procalcitonin and C-reactive protein as markers of sepsis. Crit Care Med 2003; 31: 1737-41.

9 Tang BM, Eslick GD, Craig JC, McLean AS. Accuracy of procalcitonin for sepsis diagnosis in critically ill patients: systematic review and meta-analysis. Lancet Infect Dis 2007; 7: 210-7.

10 Giamarellos-Bourboulis EJ, Mega A, Grecka P, et al. Procalcitonin: a marker to clearly differentiate systemic inflammatory response syndrome and sepsis in the critically ill patient? Intensive Care Med 2002; 28: 1351-6.

11 Endo S, Aikawa N, Fujishima S, et al. Usefulness of procalcitonin serum level for the discrimination of severe sepsis from sepsis: a multicenter prospective study. J Infect Chemother 2008; 14: 244-9.

12 Oberholzer A, Souza SM, Tschoeke SK, et al. Plasma cytokine measurements augment prognostic scores as indicators of outcome in patients with severe sepsis. Shock 2005; 23: 488-93.

13 Ugarte H, Silva E, Mercan D, De Mendonça A, Vincent JL. Procalcitonin used as a marker of infection in the intensive care unit. Crit Care Med 1999; 27: 498-504.

14 Claeys R, Vinken S, Spapen H, et al. Plasma procalcitonin and C-reactive protein in acute septic shock: clinical and biological correlates. Crit Care Med 2002; 30: 757-62.

15 Dahaba AA, Hagara B, Fall A, Rehak PH, List WF, Metzler H. Procalcitonin for early prediction of survival outcome in postoperative critically ill patients with severe sepsis. Br J Anaesth 2006; 97: 503-8.

16 Phua J, Koay ES, Lee KH. Lactate, procalcitonin, and amino-terminal pro-B-type natriuretic peptide versus cytokine measurements and clinical severity scores for prognostication in septic shock. Shock 2008; 29: 328-33.

17 Boussekey N, Leroy O, Georges H, Devos P, d’Escrivan T, Guery B. Diagnostic and prognostic values of admission procalcitonin levels in community-acquired pneumonia in an intensive care unit. Infection 2005; 33: 257-63.

18 Bell K, Wattie M, Byth K, et al. Procalcitonin: a marker of bacteraemia in SIRS. Anaesth Intensive Care 2003; 31: 629-36.

19 Clec’h C, Fosse JP, Karoubi P, Vincent F, Chouahi I, Hamza L, Cupa M, Cohen Y. Differential diagnostic value of procalcitonin in surgical and medical patients with septic shock. Crit Care Med 2006; 34: 102-7.

20 Hausfater P, Juillien G, Madonna-Py B, Haroche J, Bernard M, Riou B. Serum procalcitonin measurement as diagnostic and prognostic marker in febrile adult patients presenting to the emergency department. Crit Care 2007; 11: R60.

21 Heper Y, Akalin EH, Mistik R, et al. Evaluation of serum C-reactive protein, procalcitonin, tumor necrosis factor alpha, and interleukin-10 levels as diagnostic and prognostic parameters in patients with community-acquired sepsis, severe sepsis, and septic shock. Eur J Clin Microbiol Infect Dis 2006; 25: 481-91.

22 Pettilä V, Hynninen M, Takkunen O, Kuusela P, Valtonen M. Predictive value of procalcitonin and interleukin 6 in critically ill patients with suspected sepsis. Intensive Care Med 2002; 28: 1220-5.

23 Wunder C, Eichelbrönner O, Roewer N. Are IL-6, IL-10 and PCT plasma concentrations reliable for outcome prediction in severe sepsis? A comparison with APACHE III and SAPS II. Inflamm Res 2004; 53: 158-63.

24 Novotny A, Emmanuel K, Matevossian E, et al. Use of procalcitonin for early prediction of lethal outcome of postoperative sepsis. Am J Surg 2007; 194: 35-9.

25 Lee CC, Chen SY, Tsai CL, et al. Prognostic value of mortality in emergency department sepsis score, procalcitonin, and C-reactive protein in patients with sepsis at the emergency department. Shock 2008; 29: 322-7.

26 Viallon A, Guyomarc’h S, Marjollet O, et al. Can emergency physicians identify a high mortality subgroup of patients with sepsis: role of procalcitonin. Eur J Emerg Med 2008; 15: 26-33.

27 Loisa P, Rinne T, Laine S, Hurme M, Kaukinen S. Anti-inflammatory cytokine response and the development of multiple organ failure in severe sepsis. Acta Anaesthesiol Scand 2003; 47: 319-25.

28 Ferreira FL, Bota DP, Bross A, Mélot C, Vincent JL. Serial evaluation of the SOFA score to predict outcome in critically ill patients. JAMA 2001; 286: 1754-8.

29 Tormo A, Santolaria F, González-Reimers F, et al. Short-term prognostic value of subjective nutritional assessment in general medical patients. J Nutr Med 1994; 4: 287-95.

30 Holm A, Pedersen SS, Nexoe J, et al. Procalcitonin versus C-reactive protein for predicting pneumonia in adults with lower respiratory tract infection in primary care. Br J Gen Pract 2007; 57: 555-60.

31 Stolz D, Christ-Crain M, Gencay MM, et al. Diagnostic value of signs, symptoms and laboratory values in lower respiratory tract infection. Swiss Med Wkly 2006; 136: 434-40.

32 Routsi C, Pratikaki M, Sotiropoulou C, et al. Application of the sequential organ failure assessment (SOFA) score to bacteremic ICU patients. Infection 2007; 35: 240-4.

33 Jones AE, Fiechtl JF, Brown MD, Ballew JJ, Kline JA. Procalcitonin test in the diagnosis of bacteremia: a meta-analysis. Ann Emerg Med 2007; 50: 34-41.

34 Charles PE, Ladoire S, Aho S, et al. Serum procalcitonin elevation in critically ill patients at the onset of bacteremia caused by either Gram negative or Gram positive bacteria. BMC Infect Dis 2008; 26: 38.

35 Fraunberger P, Wang Y, Holler E, et al. Prognostic value of interleukin 6, procalcitonin, and C-reactive protein levels in intensive care unit patients during first increase of fever. Shock 2006; 26: 10-2.

36 Meng FS, Su L, Tang YQ, Wen Q, Liu YS, Liu ZF. Serum procalcitonin at the time of admission to the ICU as a predictor of short-term mortality. Clin Biochem 2009; 42: 1025-31.

37 Ruiz-Alvarez MJ, García-Valdecasas S, De Pablo R, et al. Diagnostic efficacy and prognostic value of serum procalcitonin concentration in patients with suspected sepsis. J Intensive Care Med 2009; 24: 63-71.

38 Gibot S, Cravoisy A, Kolopp-Sarda MN, et al. Time-course of sTREM (soluble triggering receptor expressed on myeloid cells)-1, procalcitonin, and C-reactive protein plasma concentrations during sepsis. Crit Care Med 2005; 33: 792-6.

39 Burgmann H, Winkler S, Locker GJ, et al. Increased serum concentration of soluble CD14 is a prognostic marker in gram-positive sepsis. Clin Immunol Immunopathol 1996; 80: 307-10.

40 Landmann R, Zimmerli W, Sansano S, et al. Increased circulating soluble CD14 is associated with high mortality in gram-negative septic shock. J Infect Dis 1995; 171: 639-44.

41 Tejera A, Santolaria F, Diez ML, et al. Prognosis of community acquired pneumonia (CAP): value of triggering receptor expressed on myeloid cells-1 (TREM-1) and other mediators of the inflammatory response. Cytokine 2007; 38: 117-23.

42 Gaïni S, Koldkjaer OG, Møller HJ, Pedersen C, Pedersen SS. A comparison of high-mobility group-box 1 protein, lipopolysaccharide-binding protein and procalcitonin in severe community-acquired infections and bacteraemia: a prospective study. Crit Care 2007; 11: R76.

43 Opal SM, Scannon PJ, Vincent JL, et al. Relationship between plasma levels of lipopolysaccharide (LPS) and LPS-binding protein in patients with severe sepsis and septic shock. J Infect Dis 1999; 180: 1584-9.

44 Sakr Y, Burgett U, Nacul FE, Reinhart K, Brunkhorst F. Lipopolysaccharide binding protein in a surgical intensive care unit: a marker of sepsis? Crit Care Med 2008; 36: 2014-22.

45 Maruna P, Gurlich R, Frasko R, Haluzik M. Serum leptin levels in septic men correlate well with C-reactive protein (CRP) and TNF-alpha but not with BMI. Physiol Res 2001; 50: 89-94.

46 Aleman MR, Santolaria F, Batista N, et al. Leptin role in advanced lung cancer. A mediator of the acute phase response or a marker of the status of nutrition? Cytokine 2002; 19: 21-6.

47 Díez ML, Santolaria F, Tejera A, et al. Serum leptin levels in community acquired pneumonia (CAP) are related to nutritional status and not to acute phase reaction. Cytokine 2008; 42: 156-60.