Lupus. Serum procalcitonin has negative predictive value for bacterial infection in active systemic lupus erythematosus

Transcription

1 in active systemic lupus erythematosus KM Bador, S Intan, S Hussin and AHA Gafor 202 2: 72 originally published online 30 May 202 DOI: 0.77/ The online version of this article can be found at: Published by: Additional services and information for can be found at: Alerts: Subscriptions: Reprints: Permissions: >> Version of Record - Sep 6, 202 OnlineFirst Version of Record - May 30, 202 What is This?

2 (202) 2, PAPER Serum procalcitonin has negative predictive value for bacterial infection in active systemic lupus erythematosus KM Bador, S Intan 2, S Hussin 3 and AHA Gafor 4 Department of Pathology, Faculty of Medicine, Universiti Kebangsaan Malaysia, Kuala Lumpur, Malaysia; 2 Department of Pathology, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, Serdang, Malaysia; 3 Department of Microbiology, Faculty of Medicine, Universiti Kebangsaan Malaysia, Kuala Lumpur, Malaysia; and 4 Department of Medicine, Faculty of Medicine, Universiti Kebangsaan Malaysia, Kuala Lumpur, Malaysia Background: Previous studies in systemic lupus erythematosus (SLE) patients have produced conflicting results regarding the diagnostic utility of procalcitonin (PCT). The aim of this study was to determine predictive values of PCT and C-reactive protein (CRP) for bacterial infection in SLE patients. Materials and methods: This was a cross-sectional study of clinic and hospitalized SLE patients with and without bacterial infection recruited over 8 months. Bacterial infection was defined as positive culture results. SLE disease activity was measured using SLEDAI. PCT and CRP were measured by automated immunoassays. Results: Sixty-eight patients (57 females) were studied. Ten patients (5%) had infection. The areas under the receiver operating characteristic curves for PCT and CRP were not significantly different [0.797 (CI ) versus (CI )]. In lupus flare patients, PCT but not CRP was higher with infection (p ¼ 0.09 versus 0.95). A PCT of <0.7 ng/ml ruled out infection with 94% negative predictive value (NPV). In remission patients, CRP but not PCT was elevated with infection (p ¼ versus 0.03). CRP < 0.57 mg/dl had 96% NPV. Conclusion: PCT may be a better marker to rule out bacterial infection in lupus flare but not in remission or general screening. (202) 2, Key words: SLE; procalcitonin; bacterial infection; predictive values Introduction Despite advances in treatment, infection remains a common cause of death in systemic lupus erythematosus (SLE) patients. Identifying infection in SLE patients can be a clinical challenge particularly during active disease (flare). Current laboratory tests such as erythrocyte sedimentation rate and leucocyte count lack accuracy to distinguish infection. A normal serum C-reactive protein (CRP) level has been suggested to indicate flare rather than infection. 2 However, both false positive and false negative CRP values have been reported to occur not uncommonly. 3,4 Microbiological Correspondence to: KM Bador, Department of Pathology, Faculty of Medicine, Universiti Kebangsaan Malaysia Medical Centre, Jalan Yaacob Latif, Cheras, Kuala Lumpur, Malaysia. khalidah@ppukm.ukm.my Received 2 September 20; accepted 4 May 202 investigations remain the confirmatory procedure but are laborious and time-consuming. Procalcitonin (PCT), a precursor of calcitonin, has been reported to be increased in non-viral infections and levels correlate with severity of illness. 5,6 Studies have shown that PCT is a useful marker of sepsis. 9,0 However, it is also raised in several noninfective inflammatory conditions. 3 In normal individuals, PCT levels are less than 0.05 ng/ml but many authors favour a higher cut-off value of 0.5 ng/ml to diagnose systemic infection. 7,8 A few studies have assessed the clinical usefulness of PCT in SLE. Leonhardt et al. studied 0 patients with systemic autoimmune disease including SLE patients and reported significant differences in PCT levels during systemic bacterial inflammation. 4 In a study by Shin et al. conducted in 9 febrile SLE patients, PCT increased significantly in those with non-viral infection compared with lupus flare and levels decreased with! The Author(s), 202. Reprints and permissions: /

3 defervescence. 5 Two studies reported that PCT was a better marker than CRP to distinguish between bacterial infection and active disease in febrile SLE patients. 6,7 However, in a retrospective study of 60 hospitalized SLE patients, Lanoix et al. concluded that PCT was not useful to differentiate patients with or without bacterial infection. 8 Previous studies have largely focused on febrile SLE patients and only one study recommended a cut-off value for PCT. 7 The aim of the present study was to compare the predictive values of PCT and CRP for bacterial infection in SLE patients and to define appropriate cut-off values. A further objective was to determine the influence of disease activity on diagnostic performances. Materials and methods Subjects A cross-sectional study was performed on patients who fulfilled the American College of Rheumatology (ACR) criteria for diagnosis of SLE. 9 SLE patients who attended clinic for follow-up consultation or who were admitted to the medical ward in the institute s medical centre were recruited consecutively during an 8-month period. Disease activity was measured using the SLE Disease Activity Index (SLEDAI) score. 20 flare was defined as an increase of more than three points compared to the patient s previous score. Patients were excluded if they were already on antibiotics or had acute myocardial infarction, acute pancreatitis, thyroid or bronchial carcinoma or a <7 days history of severe trauma or surgery. Data regarding age, gender and years since diagnosis of SLE were obtained during recruitment. The study was approved by the institute s research ethics committee and all recruited patients gave their informed written consent. Laboratory analyses Blood culture specimens were taken from all subjects immediately after clinic consultation or within 36 hours of ward admission. Urine, stool or other cultures were taken as needed to confirm the presence of infection. Bacterial infection was defined as a positive culture of aerobic or anaerobic microorganisms with identification of the bacterial species. PCT was measured by time-resolved fluorescence immunoassay on an automated platform (Brahms Diagnostica, Germany) with a limit of detection of 0.02 ng/ml. The reference range for PCT in healthy subjects is <0.05 ng/ml in our laboratory. CRP was measured by an immunoturbidimetric assay (Roche Diagnostics, Germany) with levels <0.5 mg/dl defined as the reference range for healthy subjects. Inter-batch precision of PCT and CRP assays were 6 5% and % respectively. Statistical analysis Data analyses were performed using the Statistical Package for Social Sciences (IBM Corporation, USA). Differences between groups were tested by Mann-Whitney test (2 variables) or Kruskal-Willis test (>2 variables). Spearman s correlation was used to test for associations. All tests were twotailed and a p value of <0.05 was considered significant. Receiver operating characteristic (ROC) curves were plotted to compare the sensitivity and specificity of PCT and CRP to detect bacterial infection with a 95% confidence interval (CI) of the area under the ROC curve. Results Patient characteristics Demographic and clinical characteristics of the study patients are presented in Table. A total of 68 SLE patients, mostly female (84%), were recruited. Ten patients (5%) had positive cultures for bacteria, six of whom also had lupus flare. Among the non-infected patients, 24 had flare. There was no difference in gender distribution, age (p ¼ 0.290) or years since diagnosis of SLE (p ¼ 0.087) between infected and non-infected groups of patients. Microbiological investigations identified six different bacterial species. Five patients had local infection, four patients had systemic infection (Salmonella, CoNS, chlamydia pneumonia and miliary tuberculosis) and one patient had infection of the lower limb but was clinically diagnosed and treated as having sepsis. PCT and CRP to diagnose infection in all SLE patients Both PCT (0.9 versus 0.06 ng/ml, p ¼ 0.003) and CRP (2.45 versus 0.40 mg/dl, p ¼ 0.00) were significantly higher in patients with bacterial infection compared to those without infection. PCT was positively correlated with CRP levels (r ¼ 0.39, p ¼ 0.00) but not with age (p ¼ 0.88) or duration of SLE (p ¼ 0.925). 73

4 74 Table Demographic data and clinical characteristics of SLE patients with and without bacterial infection Infected patients(n ¼ 0) Non-infected patients(n ¼ 58) p value Females 8 (80%) 49 (84%) Age (years) 32.5 (8 46) 28.5 (4 59) Years since diagnosis 8.5 (0.3 6) 4 (0 8) SLEDAI score Flare Remission Culture results Escherichia coli Enterobacter b lactamase Salmonella spp CoNS Mycobacterium tuberculosis Chlamydia pneumonia 2 (4 8) 4 (4 4) 5 8 (4 20) 0 (0 8) PCT (ng/ml) 0.9 ( ) 0.06 ( ) CRP (mg/dl) 2.45 ( ) 0.39 ( ) 0.00 Values are the median (range) except for females. CoNS: Coagulase negative staphylococcus aureus. ROC curve analysis generated similar areas under the curve for PCT (0.797; CI ) and CRP (0.755; CI ), Figure. A cut-off value of 0.2 ng/ml for PCT gave the best combination of sensitivity (80%) and specificity (78%) with positive predictive value (PPV) of 38% and negative predictive value (NPV) of 96%. Increasing the PCT threshold to 0.5 ng/ml resulted in 00% PPV (NPV 88%) but sensitivity was greatly reduced to 20%. A cut-off level of 0.48 mg/dl for CRP gave a combination of 80% sensitivity and 55% specificity with PPV and NPV of 23% and 93% respectively. To achieve 00% PPV the CRP threshold would need to be increased to 3 mg/dl. PCT and CRP levels in sub-group patients In SLE patients without infection, PCT and CRP were higher in those with active disease (p < 0.00 and p < respectively), Table 2. However, there was no correlation between SLEDAI score and PCT (r ¼ 0.58, p ¼ 0.236) or CRP levels (r ¼ 0.235, p ¼ 0.075). Among infected patients there was no significant difference in PCT (p ¼ 0.088) or CRP (p ¼ 0.83) levels between flare and remission patients. Furthermore, there was no difference between flare patients without infection and remission patients with infection (PCT, p ¼ and CRP, p ¼ 0.470). Analysis of flare patients only, showed that PCT (0.33 versus 0.08 ng/ml, p ¼ 0.09) but not CRP (3.20 versus 0.77 mg/dl, p ¼ 0.95) was higher with infection. Conversely, among remission patients, Figure ROC curves comparing sensitivity and specificity of PCT and CRP for diagnosis of bacterial infection in SLE patients (58 non-infected and 0 infected). CRP (.69 versus 0.8 mg/dl, p ¼ 0.036) but not PCT (0.4 versus 0.05 ng/ml, p ¼ 0.03) was higher with infection, Figure 2. A PCT cut-off value of 0.7 ng/ml gave the best combination of sensitivity (83%) and specificity (7%) to detect infection in lupus flare patients with NPV of 94% and PPV of 42%. A CRP cut-off value of 0.57 mg/ dl gave the best combined sensitivity (75%) and specificity (74%) to detect infection in remission patients with NPV and PPV of 96% and 25% respectively.

5 Table 2 PCT and CRP values in sub-groups of SLE patients 75 Infected patients Non-infected patients Flare(n ¼ 6) Remission(n ¼ 4) Flare(n ¼ 24) Remission(n ¼ 34) PCT (ng/ml) 0.33 ( ) 0.4 ( ) 0.08 ( ) 0.05 ( ) CRP (mg/dl) 3.20 ( ).69 ( ) 0.77 ( ) 0.8 ( ) Values are the median (range). Figure 2 Distribution of PCT and CRP levels in flare and remission patients with and without bacterial infection. Open boxes-infection, solid boxes-no infection. Boxes indicate inter-quartile range with internal horizontal line indicating median. * and o markers indicate outliers. Discussion In this study we compared the sensitivity, specificity and predictive values of PCT and CRP to diagnose bacterial infection in SLE patients presenting to a tertiary hospital. Based on the areas under the ROC curves, PCT and CRP performed equally well to differentiate bacterial infection from no infection. The cut-off values that gave the best combination of sensitivity and specificity, were 0.2 ng/ml for PCT (more than twice the reference range) and 0.48 mg/dl for CRP (at the limit of reference range) using the laboratory methods stated. At these thresholds, the sensitivities of PCT and CRP were both 80% but PCT had better specificity (78%) than CRP (55%). Both tests generated higher negative predictive values (96% for PCT and 93% for CRP) and therefore these cut-offs should be used to rule out bacterial infection. To rule in bacterial infection, the cut-off values that gave 00% specificity were PCT > 0.5 ng/ml and CRP > 3 mg/dl. However eight out of 0 infected patients, including three with systemic infection, had PCT or CRP levels that were below these threshold values thus reducing the sensitivity of these assays considerably. Only two patients had PCT > 0.5 ng/ml; one clinically diagnosed to have sepsis (PCT 2.5 ng/ml, CRP 23 mg/dl) and the other with systemic infection due to Salmonella (PCT 0.55 ng/ml, CRP 4.2 mg/dl). The low sensitivity of PCT at a threshold value of 0.5 ng/ml was previously demonstrated in the study by Lanoix et al. using the same Brahms PCT assay. 8 They reported that PCT was less than 0.5 ug/l (or 0.5 ng/ml) in all five patients who had systemic infection. The only raised PCT level occurred in a male patient with flare. Previous reports have described the superiority of PCT over CRP measurements to detect infection during lupus flare when diagnosis is particularly challenging. 6,7 Similarly, in this study PCT was a better marker than CRP for diagnosis of bacterial infection in the presence of underlying active disease. In flare patients, a PCT cut-off level of 0.7 ng/ml differentiates infection from active

6 76 disease. This cut-off value is lower than the level of 0.5 ng/ml that is frequently reported in the literature to differentiate infection from active SLE or other autoimmune diseases. 8,6,2 Ho et al. also found a higher cut-off value of 0.74 ng/ml to predict bacterial infection in febrile SLE patients with 89.5% sensitivity and 00% specificity. 7 Our diagnostic cut-off figure is lower, firstly because it is used to rule out infection. Secondly, the median PCT was 0.33 ng/ml (range ng/ml) in flare patients with infection which was much lower compared to previous reports. Saavedra et al. obtained a median PCT of.49 ng/ml in SLE patients with severe infection including one patient with septic shock. 6 The infected group in the study by Ho et al. included patients with sepsis (median PCT 32 ng/ml) and septic shock (median PCT 56 ng/ml). In this study, most patients had local infections; only one patient had sepsis (diagnosed clinically) and none had septic shock. These patient characteristics may also have contributed to the lower cut-off value obtained in this study. In remission patients, CRP performed better than PCT to differentiate bacterial infection. A CRP value of <0.57 mg/dl ruled out infection with 96% probability. Using a cut-off limit of 0.94 mg/dl, Williams et al. reported that one third of SLE patients had normal CRP during their disease and that half the CRP results were false negatives. 4 The results from sub-group analysis suggest that PCT and CRP are more useful in specific SLE patients (flare and remission respectively) using the appropriate cut-off values. However, this strategy requires prior assessment of disease activity. Furthermore, these cut-off values are based on the use of SLEDAI score although there are other methods of assessing disease activity such as the British Isles Assessment Group (BILAG) index. Consistent with previous reports, 8,22 PCT levels did not correlate with SLEDAI score. Lanoix et al. raised the possibility of age- and gender-related PCT values. 8 No association was found between PCT and age in this sample population of SLE patients but the number of male subjects was too small for analysis. A limitation of this study was the small number of infected patients, most of whom had local infections. In a study of 9 SLE patients followed up for six years in a rheumatology clinic in Hong Kong, there were 48 episodes of major bacterial infections involving 27 patients. 23 The rate of cases with infection in our study was much lower. In this study, subjects were only tested for bacterial infections whereas SLE patients are also at risk of fungal infections which can cause raised PCT levels. Patients were classified into infected and non-infected groups based on culture testing and so the possibility of a false negative culture result exists. There were 4 patients (seven with flare and seven in remission) who were clinically suspected to have infection but cultures were negative. Five of the active SLE patients had PCT < 0.7 ng/ml (cut-off to differentiate flare from infection), one had acute gastroenteritis (PCT 0.27 ng/ml) and another was treated empirically for pulmonary tuberculosis (PCT 0.3 ng/ml). All the remission patients had PCT < 0.2 ng/ml. In cases where PCT was elevated in the absence of infection, we did not determine whether PCT changed with subsequent control of disease activity. Conclusion In this cohort of SLE patients, three out of four patients with systemic infection had PCT levels below the threshold value (0.5 ng/ml) to rule in bacterial infection. PCT was higher with infection in flare patients but not in remission patients. Negative predictive values were greater than positive predictive values suggesting that PCT may be a better tool to rule out infection. Further studies are required to evaluate the use of PCT as a diagnostic tool to rule out infection in SLE. Funding This study was supported by the Faculty of Medicine Research Fund, Universiti Kebangsaan Malaysia [grant number FF ]. Conflict of interest None declared. References Cervera R, Khamashta MA, Font J, et al. Morbidity and mortality in systemic lupus erythematosus during a 5-year period: a multicenter prospective study of,000 patients. Medicine 999; 78: ter Borg EJ, Horst G, Limburg PC, van Rijswijk MH, Kallenburg CG. C-reactive protein levels during disease exacerbation and infection in systemic lupus erythematosus: a prospective longitudinal study. J Rheumatol 990; 7:

7 3 Roy S, Tan KT. Pyrexia and normal C-reactive protein (CRP) in patients with systemic lupus erythematosus: always consider the possibility of infection in febrile patients with systemic lupus erythematosus regardless of CRP levels. Rheumatology 200; 40: Williams RC, Harmon ME, Burlingame R, Du Clos TW. Studies of serum CRP in SLE. J Rheumatol 2005; 32: Brunkhorst FM, Wegscheider K, Forycki ZF, Brunkhorst R. PCT for early diagnosis and differentiation of SIRS, sepsis, severe sepsis and septic shock. Intensive Care Med 2000; 26: S Suprin E, Camus C, Gacouin A, et al. PCT: A valuable indicator of sepsis in medical ICU? Intensive Care Med 2000; 26: Sierra R. C-reactive protein and procalcitonin as markers of infection, inflammatory response and sepsis. Clin Pulm Med 2007; 2: Delevaux I, Andre M, Colombier M, et al. Can PCT measurement help in differentiating between bacterial infection and other kinds of inflammatory processes? Ann Rheum Dis 2003; 62: Assicot M, Gendrel D, Carsin H, Raymond J, Gullbaud J, Bohuon C. High serum PCT concentrations in patients with sepsis and infection. Lancet 993; 34: Levy M, Fink M, Marshal J, et al. SCCM/ESICM/ACCP/ATS/ SIS International sepsis definitions conference. Crit Care Med 200; 3: Meisner M, Tschaikowsky K, Hutzler A, Schick C, Schuttler J. Postoperative plasma concentrations of procalcitonin after different types of surgery. Intensive Care Med 998; 24: Wanner GA, Keel M, Steckholzer U, Beier W, Stocker R, Ertel W. Relationship between procalcitonin plasma levels and severity of injury, sepsis, organ failure, and mortality in injured patients. Crit Care Med 2000; 28: Carsin H, Assicot M, Feger F, et al. Evolution and significance of circulating PCT levels compared with IL-6, TNFa and endotoxin levels early after thermal injury. Burns 997; 23: Leonhardt U, Werner M, Engelmann L. Procalcitonin in autoimmune disease with sepsis. Crit Care 200; 5: S25 (abstract P52). 5 Shin KC, Lee YJ, Kwang SW, et al. Serum PCT measurements for detection of intercurrent infection in febrile patients with SLE. Ann Rheum Dis 200; 60: 988 (letter). 6 Saavedra PG, Velasquez M, Uribe O, Vasquez G, Ramirez L. PCT in the differential diagnosis of fever in patients with SLE. Ann Rheum Dis 2006; 65: 358 (abstract FRI0238). 7 Ho W-L, Lan J-L, Chen D-Y, Chen Y-H, Huang W-N, Hsieh T-Y. Procalcitonin may be a potential biomarker for distinguishing bacterial infection from disease activity in febrile patients with systemic lupus erythematosus. Formosan Journal of Rheumatology 2009; 23: Lanoix JP, Bourgeois AM, Schmidt J, et al. Serum procalcitonin does not differentiate between infection and disease flare in patients with systemic lupus erythematosus. 200; 6. 9 Hochberg MC. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 997; 40: 725 (letter). 20 Urowitz MB, Gladman DD. Measures of disease activity and damage in SLE. Clin Rheumatol 998; 2: Eberhard OK, Haubitz M, Brunkhorst FM, Kliem V, Koch KM, Brunkhorst R. Usefulness of PCT for differentiation between activity of systemic autoimmune disease (SLE/systemic antineutrophil cytoplasmic antibody associated vasculitis) and invasive bacterial infection. Arthritis Rheum 997; 40: Quintana G, Medina Y, Rojas C, et al. The use of procalcitonin determinations in evaluation of systemic lupus erythematosus. J Clin Rheum 2008; 4: Ng WL, Chu CM, Wu AKL, Cheng VCC, Yuen KY. Lymphopenia at presentation is associated with increased risk of infections among patients with systemic lupus erythematosus. QJM 2006; 99: