Original Studies

Predictive Value of Cerebrospinal Fluid Matrix Metalloproteinase-9 and Tissue Inhibitor of Metalloproteinase-1 Concentrations in Childhood Bacterial Meningitis Irmeli Roine, MD, PhD,* Tuula Pelkonen, MD, PhD,†‡ Luis Bernardino, MD,† Anneli Lauhio, MD, PhD,§ Taina Tervahartiala, DDS, PhD,¶ Maija Lappalainen, MD, PhD,‖ Matti Kataja, PhD,** Anne Pitkäranta, MD, PhD,††‡‡ Timo Sorsa, DDS, PhD,¶§§ and Heikki Peltola, MD, PhD‡ Background: Increased concentrations of matrix metalloproteinases (MMP) in cerebrospinal fluid are part of the host response in bacterial meningitis (BM). We investigated whether the concentrations of MMP-9 and the tissue inhibitor of metalloproteinase (TIMP)-1 predict the outcome in childhood BM. Methods: Cerebrospinal fluid MMP-9 and tissue inhibitor of ­metalloproteinase-1 (TIMP-1) were quantified by an enzyme-linked immunosorbent assay from 264 and 335 patients, respectively; 43 children without BM served as controls. The results were compared with previously known independent predictors of death and sequelae. Results: Higher MMP-9 and TIMP-1 values distinguished the controls from the BM patients (P < 0.0001). A MMP-9 concentration >940 ng/mL proved an independent predictor of death [adjusted odds ratio: 4.03; 95% confidence interval (CI): 2.09−7.77; P < 0.0001]. If the patient additionally presented with a Glasgow Coma Score below 9, the odds increased to 13.21 (95% CI: 5.44−32.08; P < 0.0001). TIMP-1 levels correlated with the severity of sequelae (ρ: 0.30; P < 0.0001), but not with death. Its concentration above 390 ng/mL increased the likelihood of sequelae 3.43fold (95% CI: 1∙73−6∙79; P = 0.0004), and up to 31.18-fold (95% CI: 4.05−239.8; P = 0.0009) if the patient also presented a Glasgow Coma Score < 12. Conclusions: Elevated cerebrospinal fluid MMP-9 and TIMP-1 values predict 2 important outcomes in childhood BM. Combined with a clinical evaluation, quantification of these indices augments the chances to identify the patients in greatest need of better treatment modalities.

B

acterial meningitis (BM) is a life threatening disease, which can lead to irreparable injury of the cognitive, motor and sensory functions of the central nervous system.1 Its burden remains high in areas where vaccines are underused; the estimated world incidence exceeds 1 million cases per year.2,3 Adverse outcomes are associated with the severity of disease on admission and relate with the extent of inflammation in the central nervous system.4,5 An increase in cerebrospinal fluid (CSF) concentrations of matrix metalloproteinase (MMP)-8 and MMP-9 and tissue inhibitor of metalloproteinase (TIMP)-1 forms a part of the host response in BM.6–8 MMP-9 correlates with the TNF-α level,6 blood brain barrier damage9 and the CSF white cell count.7,10,11 A reduction of the enzymatic activity of MMP-9 reduced the inflammatory response in experimental BM and improved prognosis.12–14 This effect was independent of the timing of the administration of the antibiotic,14 in contrast to dexamethasone.15,16 In a small study (n = 27) of childhood BM, increased CSF MMP-9 levels have been associated with permanent neurologic sequelae.6 The ratio of MMP-9/TIMP-1, a measure of MMP-9´s proteolytic activity,6 may be even more informative of outcome than either alone. We measured the CSF concentrations of MMP-9 and TIMP-1 and their molar ratio in a large number of children with BM to assess their relation with other parameters that used to grade meningitis severity and to test the extent of their association with the outcome.

Key Words: matrix metalloproteinase-9, tissue inhibitor of metalloproteinase, bacterial meningitis, outcome, childhood

MATERIALS AND METHODS

(Pediatr Infect Dis J 2014;33:675–679)

Patients

Accepted for publication November 25, 2013. From the *Faculty of Medicine, University Diego Portales, Santiago, Chile; †David Bernardino Children’s Hospital, Luanda, Angola; ‡Children´s Hospital; §Division of Infectious Diseases, Department of Medicine; ¶Department of Oral and Maxillofacial Diseases; ‖Department of Virology and Immunology, Laboratory Services (HUSLAB), Helsinki University Central Hospital; **National Institute for Health and Welfare; ††Institute of Clinical Medicine; ‡‡Department of Otorhinolaryngology, Helsinki University Central Hospital; and §§Institute of Dentistry, University of Helsinki, Helsinki, Finland. This work was supported by grants from the Päivikki and Sakari Sohlberg Foundation, the Sigrid Jusélius Foundation and the Foundation for Pediatric Research, Finland. The authors have no conflicts of interest to disclose. Address for correspondence: Irmeli Roine, MD, PhD, Los Misioneros 2237, 7520179 Santiago, Chile. E-mail: [email protected]. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (www.pidj.com). Copyright © 2013 by Lippincott Williams & Wilkins ISSN: 0891-3668/14/3307-0675 DOI: 10.1097/INF.0000000000000249

The present study is part of a previously published prospective, randomized trial on 723 children with BM, which examined the effects of β-lactam infusion and high-dose oral paracetamol in the Paediatric Hospital of Luanda, Angola.17 The study was registered (ISRCTN62824827) and approved by the Children’s Hospital’s Ethics Committee. The children were enrolled with the guardians’ informed consent. The Ethics Committee declined the use of adjuvant corticosteroids; instead, all patients received oral glycerol. The present analysis included the children with confirmed BM (n = 553/723), of whom a large enough pretreatment CSF sample was available for the MMP-9 (n = 264) and/or the TIMP-1 (n = 335) measurement. At discharge from hospital, the survivors were categorized into 2 groups, those without sequelae and those with some neurologic or audiologic sequelae. The predetermined criteria for severe (major) neurologic sequelae were blindness, quadriplegia or paresis, hydrocephalus requiring a shunt or severe psychomotor retardation (according to Denver Developmental Screening Test). Minor neurologic sequelae comprised moderate psychomotor retardation, hemiplegia or paresis, monoparesis or ataxia. Hearing was tested

The Pediatric Infectious Disease Journal  •  Volume 33, Number 7, July 2014

www.pidj.com | 675

The Pediatric Infectious Disease Journal  •  Volume 33, Number 7, July 2014

Roine et al

by brainstem auditory evoked potentials (Madsen Octavus™, GN Otometrics A/S, Taastrup, Denmark) and classified as normal when the hearing threshold of the better ear was 40 dB, moderately impaired when 60 dB and severely impaired when ≥80 dB.

Samples CSF samples were obtained at study entry, stored in ­ eep-freeze and transported frozen to Finland. The control specid mens were anonymous pool samples from 43 children attended at the Children’s Hospital, Helsinki University Central Hospital for suspected but unconfirmed bacterial infection of the central nervous system. The MMP-9 and TIMP-1 concentrations were determined by an enzyme-linked immunosorbent assay, as described previously.18 The detection limit for MMP-9 was 0.6 ng/mL and for TIMP-1 1.25 ng/mL. The coefficients of variation were 8.8% and 13.1%, respectively.

Statistical Analysis The data were analyzed with StatView (version 5.1) and are presented as medians with interquartile range (IQR) in parenthesis. Mann-Whitney U test (2 groups) and Kruskall-Wallis test (3 or more groups) were used to test numeral differences between the groups. Spearman rank tested the correlation of MMP-9, TIMP-1 and their molar ratio with other quantitative variables. A logistic regression model was used to test the predictive values of MMP-9 and TIMP-1 in relation to death and sequelae and to adjust the results to other previously known independent predictors of death in the same children (altered status of consciousness, severe or moderate respiratory distress, delayed presentation and high blood glucose).19 The results are expressed as odds ratio (OR) with 95% confidence intervals (95% CI). P < 0.05 was taken as significant.

RESULTS BM Versus Control Group MMP-9 was detected (Fig. 1) in the CSF in 98% (259/264) of the BM patients and in 12% (5/43) of the controls, whereas TIMP-1 was found in 100% of both the BM (335/335) and the

control patients (42/42). The median values of MMP-9 (463 ng/mL) and TIMP-1 (330 ng/mL) and their molar ratio (0.35) in BM were higher than that in the controls (not detected, 83 ng/mL, not detected, P < 0.0001 for each, respectively; Table, Supplemental Digital Content 1, http://links.lww.com/INF/B796). These variables showed wide distributions, with outliers and some degree of overlapping between BM and the control group. Least overlapping was observed in the MMP-9 levels (Table, Supplemental Digital Content 1, http://links.lww.com/INF/B796).

Associations With Other Parameters in BM MMP-9 concentration was higher in older (ρ: 0.24; P = 0.0001, Table 1) and severely underweight children (weight/age below − 2 standard deviations, P = 0.006, Table 2) and in meningitis caused by Gram-positive bacteria (P = 0.007, Table 2). In meningitis caused by Streptococcus pneumoniae (Fig. 2), the MMP-9 level was 557 ng/mL (IQR: 946, n = 97), whereas in cases due to Haemophilus influenzae, Neisseria meningitidis and other bacteria it was 254 ng/mL (IQR: 610, n = 79), 476 ng/mL (IQR: 586, n = 24) and 1174 ng/mL (IQR: 1714, n = 13), respectively. The MMP-9 level correlated with several measures of severity. Higher levels characterized children arriving in a poor general condition (P = 0.005, Table 2), with more signs indicating increased intracranial pressure (ρ: 0.20; P = 0.001), higher plasma C-reactive protein level (ρ: 0.22; P = 0.005), higher Simple Luanda Score19 (ρ: 0.18; P = 0.004) and lower Glasgow Coma Score (GCS, ρ: −0.19; P = 0.003, Table 1). In relation to the CSF changes, increasing MMP-9 levels correlated with decreasing glucose concentrations (ρ: −0.34; P < 0.0001) and increasing white cell counts (ρ: 0.19; P = 0.002). Lesser associations were seen for systolic blood pressure (ρ: 0.16; P = 0.01) and plasma hemoglobin (ρ: 0.16; P = 0.01). No correlation was observed between MMP-9 and TIMP-1 levels, days of illness before admission, blood white cell count, HIV status or sickle cell screening positivity (P > 0.05, Table 1). In addition, the TIMP-1 levels correlated with age, but in a direction opposite to MMP-9; older children showed lower values (ρ: −0.16; P = 0.003, Table 1). In contrast to MMP-9, which showed

FIGURE 1.  CSF MMP-9 and TIMP-1 concentrations and their molar ratio in bacterial meningitis versus controls. In each box, the middle horizontal line represents the median value and the indentations mark the 95% CIs. The upper and lower horizontal lines show the 75th and 25th percentiles, respectively, whereas the vertical lines indicate the span between the 90th and 10th percentiles and the dots indicate the outliers from these limits.

676  |  www.pidj.com

© 2013 Lippincott Williams & Wilkins

The Pediatric Infectious Disease Journal  •  Volume 33, Number 7, July 2014

MMP-9 and TIMP-1 in Bacterial Meningitis

TABLE 1.  Correlations Between CSF Levels of MMP-9, TIMP-1 and Their Molar Ratio With Other Quantitative Clinical and laboratory Variables MMP-9  ng/mL Variable CSF TIMP-1 ng/mL CSF MMP-8 ng/mL Age in months Systolic blood pressure mm Hg Days of illness before admission Intracranial pressure signs, number GCS* Herson-Todd Score† Simple Luanda Score‡ CSF white cell count/μL CSF glucose mg/dL C-reactive protein mg/L Blood hemoglobin g/dL Blood white cell count/μL Days of hospital stay Glasgow Outcome Score§ at discharge

TIMP-1  ng/mL

MMP-9/TIMP-1

ρ

P value

Ρ

P value

ρ

P value

−0.03 0.38 0.24 0.16 0.003 0.20 −0.19 0.04 0.18 0.19 −0.34 0.22 0.16 0.12 0.21 −0.26

0.69

Predictive value of cerebrospinal fluid matrix metalloproteinase-9 and tissue inhibitor of metalloproteinase-1 concentrations in childhood bacterial meningitis.

Increased concentrations of matrix metalloproteinases (MMP) in cerebrospinal fluid are part of the host response in bacterial meningitis (BM). We inve...
341KB Sizes 1 Downloads 0 Views