Home > Subjects > Critical Care > Serum MMP-8 and TIMP-1 in Critically Ill Patients with Acute...
Anesthesia & Analgesia:
doi: 10.1213/ANE.0000000000000120
Critical Care, Trauma and Resuscitation: Research Report

Serum MMP-8 and TIMP-1 in Critically Ill Patients with Acute Respiratory Failure: TIMP-1 Is Associated with Increased 90-Day Mortality

Hästbacka, Johanna MD, PhD*; Linko, Rita MD, PhD*; Tervahartiala, Taina DDS, PhD; Varpula, Tero MD, PhD*; Hovilehto, Seppo MD; Parviainen, Ilkka MD, PhD§; Vaara, Suvi T. MD, PhD*; Sorsa, Timo DDS, PhD; Pettilä, Ville MD, PhD*

Free Access
Article Outline
Collapse Box

Author Information

From the *Intensive Care Units, Helsinki University Hospital; Department of Oral and Maxillofacial Diseases, Helsinki University Hospital and Biomedicum Helsinki, Helsinki; Intensive Care Unit, South Carelia Central Hospital, Lappeenranta; and §Department of Anesthesiology and Intensive Care, Kuopio University Hospital, Kuopio, Finland.

Accepted for publication December 20, 2013.

Funding: Supported by EVO-grant from Helsinki University Hospital [T102010070], [TYH 2012210] and Academy of Finland.

The authors declare no conflicts of interest.

The results of this study have been partly presented as a poster in the annual congress of the European Society of Intensive Care Medicine in Lisbon on October 16, 2012.

Reprints will not be available from the authors.

Address correspondence to Johanna Hästbacka, MD, PhD, Intensive Care Units, Helsinki University Hospital, Intensive Care Unit, Jorvi Hospital, Turuntie 150, 02740 Espoo, Finland. Address e-mail to johanna.hastbacka@hus.fi.

Collapse Box


BACKGROUND: Matrix metalloproteinases (MMPs) likely have an important role in the pathophysiology of acute lung injury. In a recent study, high matrix metalloproteinases (MMP-8) levels in tracheal aspirates of pediatric acute respiratory distress syndrome (ARDS) patients were associated with worse outcome. In patients with sepsis, an imbalance between MMPs and their tissue inhibitors (TIMPs) has been associated with impaired survival. We hypothesized that the elevated systemic MMP-8 and TIMP-1 are associated with worse outcome in acute respiratory failure.

METHODS: This was a substudy of the observational FINNALI study conducted in 25 Finnish intensive care units over an 8-week period. All patients older than 16 years requiring mechanical ventilation for >6 hours were included. MMP-8 and TIMP-1 levels were analyzed from blood samples taken on enrollment in the study and 48 hours later. Laboratory analyses were performed by using immunofluorometric assay for MMP-8 and ELISA for TIMP-1. MMP-8 and TIMP-1 levels were compared between 90-day survivors and nonsurvivors. Survival was compared in quartiles based on TIMP-1 levels, and ROC analysis was performed to calculate areas under the curves. The relationship between MMP-8 and TIMP-1 levels and degree of hypoxemia was examined.

RESULTS: The final analyses included 563 patients. Admission TIMP-1 levels were higher in nonsurvivors, median 367 ng/mL (interquartile range 199–562), than survivors, median 240 ng/mL (interquartile range 142–412), WMWodds 1.68 (95% confidence interval [CI], 1.43–2.08). MMP-8 levels may have differed between survivors and nonsurvivors, WMWodds 1.20 (95% CI, 1.01–1.43), but no difference was found in the MMP-8/TIMP-1 molar ratio, WMWodds 0.83 (95% CI, 0.67–1.04). Difference in survival between quartiles based on TIMP-1 was significant (log-rank, P < 0.001). ROC analysis produced an area under the curve 0.63 (95% CI, 0.58–0.69) for TIMP-1. TIMP-1 was associated with severity of hypoxemia. TIMP-1 levels were higher in an ARDS subgroup than in the whole cohort, WMWodds 1.65 (95% CI, 1.15–2.44).

CONCLUSIONS: MMP-8 levels were possibly higher in 90-day nonsurvivors but performed poorly in predicting outcome. Increased systemic levels of TIMP-1 were associated with more severe hypoxemia and worse outcome in a large cohort of mechanically ventilated critically ill patients and in a subgroup of ARDS patients.

Acute respiratory failure (ARF) is the leading cause of intensive care admissions.1 ARF lacks a generally accepted definition, and conditions of both pulmonary and nonpulmonary origin may lead to a need for mechanical ventilation. Patients with acute lung injury (ALI) or acute respiratory distress syndrome (ARDS) comprise a subgroup of ARF patients, with a mortality of up to 45% in severe ARDS.2 Recently, a new definition for ARDS, the Berlin Definition, was published. The Berlin Definition classifies ARDS patients into 3 severity categories. Mild ARDS corresponds to, and has now replaced, the definition of ALI.

Inflammatory variables are not included in the new definition. Nevertheless, inflammation is considered a central feature in the pathophysiology of ARDS. Leukocytes characteristically accumulate in the alveolar space in the early phase.3 The transmigration of neutrophils from pulmonary capillaries to the site of injury is facilitated by several proteinases that are stored in intracellular granules and released on the presence of proinflammatory stimuli.4 These potent proteinases include, for example, neutrophil elastase, myeloperoxidase, and matrix metalloproteinases MMP-8 and MMP-9.4

MMPs are zinc-dependent endopeptidases that are able to degrade almost all extracellular matrix components.5 Of the leukocyte-derived MMPs, MMP-8 digests type I collagen.6 In addition to participating in degradation, turnover, and remodeling of the extracellular matrix, MMPs participate in almost all stages of the inflammatory response.7 The function of MMPs is strictly regulated. The most important inhibitors at the tissue level are the tissue inhibitors of metalloproteinases (TIMPs).8 In fulminant inflammation, the inhibitory capacity of TIMPs may be overwhelmed, leading to excessive tissue damage9 and adverse outcome.

In a recent study, high MMP-8 in tracheal aspirate was associated with illness severity and increased ventilator days in pediatric patients with ARDS.10 In critically ill patients with sepsis or septic shock, an association of an imbalance between MMPs and their tissue inhibitors and increased mortality has been described,11–13 and high TIMP-1 levels have been found in nonsurvivors.11–13 Previous studies suggest that MMPs may have an important role in ALI, although this role may be either harmful or beneficial.14

We hypothesized that in ARF, and especially in ARDS, the systemic levels of MMP-8 and TIMP-1 may correlate with disease severity and patient mortality. Accordingly, we analyzed the association of serum MMP-8, TIMP-1, and the MMP-8/TIMP-1 ratio with 90-day mortality in a large prospective cohort of critically ill patients with ARF.

Back to Top | Article Outline


The local IRB of each hospital approved the study design. The requirement for written informed consent was waived for data collection, but a written informed consent was obtained from patients or their surrogates for the collection of blood samples. This was a prospective substudy of a large multicenter observational study that included all ARF patients in 25 Finnish intensive care units (ICU) during an 8-week period (FINNALI study). The study design has been described in detail elsewhere.15 Briefly, all patients older than 16 years requiring >6 hours of mechanical ventilation for ARF (either with invasive or noninvasive interface) were included. Demographic data, medications, risk factors for ARDS, and mechanical ventilation variables were recorded in a separate electronic daily case report form. Clinical diagnoses and clinical variables, severity scorings (Simplified Acute Physiology Score, SAPS II), daily organ dysfunction scorings (Sequential Organ Failure Assessment, SOFA), and ICU mortality were derived from the routine intensive care dataset (Finnish Quality Consortium, Tieto Oy, Helsinki, Finland). In addition, 30-day and 90-day mortality data were obtained from Statistics Finland.

ALI/ARDS was defined according to the American-European Consensus Conference (AECC) criteria by the clinicians.16 In accord with the Berlin Definition,2 the term ALI is avoided in the text. Patients fulfilling the former ALI criteria are referred to as ARDS patients. Patients with sepsis with a vasopressor therapy on admission were classified as septic shock patients. Patients classified as immunocompromised and those receiving systemic corticosteroids or cytostatic medication were excluded from this substudy. Patients admitted to the ICU after an elective operation are referred to in the text as elective patients. This group includes patients who underwent cardiothoracic or major abdominal surgery and other electively operated patients who, due to a severe underlying disease, received planned postoperative care in an ICU.

Back to Top | Article Outline
Blood Samples

Serum samples for MMP-8 and TIMP-1 were obtained from an arterial line or venipuncture at study inclusion (A) (= 6 hours after initiating mechanical ventilation) and at 48 hours after inclusion (B). The sample with a higher enzyme concentration of these 2 samples (A or B), when both samples were available, is referred to in the text as H (MMP-8H, TIMP-1H). The samples were centrifuged for 15 minutes at 2000g, and the supernatant was then divided into aliquots and stored at a minimum of −20°C until analysis.

Back to Top | Article Outline
Laboratory Analysis of MMP-8 and TIMP-1

The serum levels of MMP-8 were analyzed by a time-resolved immunofluorometric assay (IFMA). The monoclonal MMP-8 specific antibodies 8708 and 8706 (Medix Biochemica, Kauniainen, Finland) were used as a catching antibody and a tracing antibody, respectively. The concentrations of MMP-8 are expressed in ng/mL, and the detection limit was 0.08 ng/mL.17,18 TIMP-1 analyses were performed by using commercial enzyme-linked immunosorbent assay (ELISA) kits (Biotrak ELISA System, Amersham Biosciences, Buckinghamshire, UK).The detection limit of the kit for TIMP-1 is 1.25 ng/mL.

The samples were assayed in random order according to the manufacturer’s instructions. The concentrations of TIMP-1 are expressed in ng/mL.

Back to Top | Article Outline
Statistical Analysis

The primary end point was 90-day mortality. Considering baseline characteristics, continuous data of independent groups were compared by using a 2-tailed Mann-Whitney test. For comparisons of the nonparametric main variables between independent groups, we calculated WMWodds with 95% confidence intervals (CIs) to assess associations with outcome or diagnostic group.19 Categorical data were compared by using Fisher exact test. Due to the nonnormal distribution of the laboratory parameters, we used the Spearman nonparametric correlation to test relationships. We constructed receiver operating characteristic (ROC) curves and calculated areas under the curve (AUC) with 95% CIs to assess predictive powers of MMP-8A, TIMP-1A, and TIMP-1H regarding 90-day mortality. To test the ability of a variable to independently predict outcome, we used backward logistic regression analysis. Interactions between covariates were tested, and if a significant interaction was present, we included interaction term in the final model. We plotted the logit of the probability for 90-day death, obtained from the model, against each independent covariate included in the final step of the model to test the linearity assumption of logistic regression. The goodness of fit was assessed with Hosmer-Lemeshow statistics. We defined cutoff values with 95% CIs by using the Youden20 method and then calculated positive likelihood ratios with 95% CIs for the cutoff values in predicting mortality. Comparisons between continuous nonparametric data in several groups were performed by using the Kruskal-Wallis test. Finally, we constructed Kaplan-Meier survival curves according to quartiles of MMP-8 and TIMP-1 to demonstrate the difference in survival between groups.

The statistical analyses were performed by using SPSS statistics 20.0 (IBM SPSS, Chicago, IL) and MedCalc® Software (MedCalc®, Osteend, Belgium). The data are presented as median and interquartile range (IQR), absolute values and percentages, or means (SD). In all comparisons, P < 0.01 was considered significant due to multiple comparisons.

Back to Top | Article Outline


After excluding immunocompromised patients (N =61) and patients with no blood samples on admission and/or 48 hours after admission, 563 patients were included in the final analyses. The study flowchart is presented in Figure 1. Patient characteristics, severity scores, and outcome data for all patients and the ARDS subgroup are shown in Table 1.

Figure 1
Figure 1
Image Tools
Table 1
Table 1
Image Tools

Median MMP-8 concentrations on admission were 96 ng/mL (IQR 44–188 ng/mL) and 147 ng/mL (IQR 53–214 ng/mL) for the whole cohort and the subgroup, respectively. Corresponding median TIMP-1 concentrations were 265 ng/mL (IQR 154–466 ng/mL) and 348 ng/mL (IQR 214–663 ng/mL), respectively.

Back to Top | Article Outline
Association of MMP-8 and TIMP-1 with 90-Day Mortality

Of the 563 patients, 132 (23.4%) died within 30 and 152 (27.0%) within 90 days of hospital admission. ICU mortality was 9.2% (N = 52). A comparison of demographics, severity scores, and admission MMP-8, TIMP-1, and MMP-8/TIMP-1 molar ratio between 90-day survivors and nonsurvivors is shown in Table 2.

Table 2
Table 2
Image Tools

The 90-day nonsurvivors were older and had more severe disease on admission by SAPS II, SOFA at 24 hours from admission, and maximal SOFA score. Among the 90-day nonsurvivors, there was significantly more septic shock, chronic cardiac disease, ARDS, and fewer postoperative patients than among 90-day survivors. The groups were not different with respect to the presence of infection or sepsis. A significant difference emerged in admission TIMP-1 levels, WMWodds 1.68 (95% CI, 1.43–2.08), and possibly in admission MMP-8, WMWodds 1.20 (95% CI, 1.01–1.43) but not in the MMP-8/TIMP-1 molar ratio, WMWodds 0.83 (95% CI, 0.67–1.04), between 90-day survivors and nonsurvivors.

To test the ability of MMP-8 and TIMP-1 to differentiate survivors from nonsurvivors, we constructed ROC curves and calculated AUCs for MMP-8A and TIMP-1A and TIMP-1H. Figure 2 demonstrates the ROC curves for MMP-8A and TIMP-1A. The AUCs for MMP-8A, TIMP-1A, and TIMP-1H were 0.55 (95% CI, 0.49–0.60), 0.63 (95% CI, 0.58–0.69), and 0.61 (95% CI, 0.55–0.67), respectively. In backward logistic regression analysis, TIMP-1A was independently associated (P = 0.004) with 90-day mortality when MMP-8, age, SAPS II without age points, SOFA at 24 hours from admission, and chronic cardiac disease were included in the model. The log odds were linearly related to each independent variable. The model also included interaction terms between (1) SAPS II without age points and age and (2) SOFA score and age. The Hosmer-Lemeshow χ2 of the final model was 11.14, P = 0.194.

Figure 2
Figure 2
Image Tools

At the optimal cutoff level of TIMP-1A 458.6 ng/mL (95% CI, 362.1–736.7 ng/mL), the sensitivity and specificity of the test to discriminate 90-day nonsurvivors and 90-day survivors were 0.41 and 0.80, respectively. The positive likelihood ratios for TIMP-1A and TIMP-1H at the cutoff level of 458.6 ng/mL were 2.08 (95% CI, 1.6–2.8) and 1.63 (95% CI, 1.25–2.13), respectively. At the cutoff level of 127.3 ng/mL (95% CI, 31.3–330.5 ng/mL), MMP-8A had a positive likelihood ratio of 1.26 (95% CI, 1.00–1.60).

To further investigate the association of TIMP-1 with mortality, we calculated 90-day mortality in different quartiles of TIMP-1A. The Kaplan-Meier survival curves are shown in Figure 3. The difference in 90-day survival between groups based on TIMP-1A levels divided into quartiles was statistically significant. In patients with TIMP-1A levels exceeding 458.6 ng/mL, mortality was significantly higher (P < 0.001) than in patients with levels below the cutoff.

Figure 3
Figure 3
Image Tools

Associations of MMP-8 and TIMP-1 levels with severity of respiratory failure were significant (P = 0.002 and P< 0.001, respectively). In the quartile with the lowest PaO2/FIO2 on admission (lowest value during the preceding 6 hours of mechanical ventilation), MMP-8 (P = 0.002) and TIMP-1 (P < 0.001) levels were higher than in the highest PaO2/FIO2 quartile, and the levels decreased sequentially with increasing PaO2/FIO2 ratio.

Back to Top | Article Outline
Correlations of MMP-8 and TIMP-1 with Infection Parameters and Severity of Disease

MMP-8A correlated significantly with TIMP-1A (P < 0.001, Rho = 0.247), white blood cell (WBC) count (P < 0.001, Rho=0.241), and C-reactive protein (CRP) (P< 0.001, Rho=0.409). MMP-8A was negatively correlated with PaO2/FIO2 ratio (P < 0.001, Rho = −0.162). TIMP-1A correlated with MMP-8A, CRP (P < 0.001, Rho = 0.409), SOFA score at 24 hours from ICU admission (P < 0.001, Rho = 0.323), and SAPS II score (P < 0.001, Rho = 0.162). TIMP-1A was negatively correlated with PaO2/FIO2 ratio (P < 0.001, Rho = −0.260).

Back to Top | Article Outline
MMP-8 and TIMP-1 in Different Modes of Mechanical Ventilation

Most patients with an admission blood sample (N =426, 81.6%) received invasive mechanical ventilation during the study, whereas 63 patients (11.2%) received only noninvasive ventilatory support by either conventional continuous positive airway pressure (CPAP) or CPAP with pressure support. In 18 patients (3.4%), noninvasive ventilation was converted to invasive ventilation before inclusion in the study. In 17 patients (3.3%), noninvasive ventilation was converted to invasive ventilation after inclusion in the study. Admission MMP-8 levels were lowest in the invasively ventilated patients (P < 0.001). There were no differences in TIMP-1A levels among the 4 groups (P = 0.630). SAPS II score and SOFA score at 24 hours were higher in the invasively ventilated patients, but we detected no differences in mortality among the groups. MMP-8A was not different between survivors and nonsurvivors in any of the groups. TIMP-1A was significantly higher, WMWodds 1.85 (95% CI, 1.44–2.42), in nonsurvivors only in the group of invasively ventilated patients, which represented the largest group of patients.

Back to Top | Article Outline
MMP-8 and TIMP-1 in Patients with ARDS

The ARDS subgroup comprised 43 patients. Twenty-four (55.8%) of these patients were classified as ALI/non-ARDS by the AECC criteria. Admission MMP-8 was not different in the ARDS subgroup than in the other patients, median 147 ng/mL (IQR 53–214 ng/mL) vs 94 ng/mL (IQR 43–179 ng/mL), WMWodds 1.24 (95% CI, 0.85–1.84). TIMP-1A was higher in the ARDS subgroup than in other patients, median 348 ng/mL (IQR 214–663 ng/mL) vs 254 ng/mL (IQR 151–458 ng/mL), WMWodds 1.65 (95% CI, 1.15–2.44). In the ARDS subgroup, the differences in MMP-8A, WMWodds 0.80 (95% CI, 0.37–1.64) or TIMP-1A, WMW odds 1.88 (95% CI, 0.93–4.71), levels between 90-day survivors and nonsurvivors were not significant. AUC for TIMP-1A in ROC analysis was 0.69 (95% CI, 0.52–0.85) (P = 0.03). At the cutoff level of 392.4 ng/mL, the sensitivity and specificity for differentiating 90-day nonsurvivors from 90-day survivors were 0.62 and 0.76, respectively, and the positive likelihood ratio was 2.60 (95% CI, 1.13–6.00). Survival in groups divided into quartiles, according to TIMP-1A concentrations, may have differed among groups (log-rank, P = 0.02).

Back to Top | Article Outline
MMP-8 and TIMP-1 in Patients with Sepsis

Because MMP-8 and TIMP-1 are upregulated in septic patients,12 we performed a subgroup analysis on patients diagnosed as having sepsis within 48 hours of admission. Of the 522 patients with an admission blood sample, 68 (13%) were diagnosed as having sepsis (information missing in 2 patients).

Median MMP-8A levels in nonseptic (N = 452) and septic (N = 68) patients were 89.27 ng/mL (IQR 41.57–166.52 ng/mL) and 170.82 ng/mL (IQR 81.79–287.81 ng/mL), respectively, WMWodds 2.06 (95% CI, 1.52–2.90). The corresponding median TIMP-1A levels were 233.44 ng/mL (IQR 142.12–406.31 ng/mL) and 573.98 ng/mL (IQR 372.32–750.64 ng/mL), respectively, WMWodds 5.06 (95% CI, 3.74–7.33). MMP-8A levels did not differ significantly between 90-day survivors and nonsurvivors in either nonseptic patients, WMWodds 1.19 (95% CI, 0.93–1.53) or septic patients, WMWodds 0.88 (95% CI, 0.49–1.57). TIMP-1A levels were higher in 90-day nonsurvivors than in 90-day survivors in patients with no diagnosed sepsis, WMWodds 1.67 (95% CI, 1.31–2.16). In septic patients, the difference was not significant, WMWodds 1.60 (95% CI, 0.89–3.18).

Back to Top | Article Outline
MMP-8 and TIMP-1 in Elective and Emergency Admissions

Ninety-five patients (17%) were admitted to an ICU after an elective operation. Of these, 59 (62%) were operated on for cardiac disease, 7 (7%) for respiratory neoplasms, 7 (7%) for gastrointestinal neoplasms, 6 (6%) for abdominal aortic aneurysms, and 16 (17%) for other diseases. To control for possible bias from this patient group, we analyzed TIMP-1A and MMP-8A separately in these 2 groups. Median MMP-8A and TIMP-1A in elective patients were 61 ng/mL (IQR 37–120 ng/mL) and 254 ng/mL (IQR 164–449 ng/mL), respectively. In patients with emergency admission, median MMP-8A and TIMP-1A were 105 ng/mL (IQR 46–205 ng/mL) and 269 ng/mL (IQR 144–472 ng/mL), respectively. The difference between these groups was significant in MMP-8A, WMWodds 1.54 (95% CI, 1.21–1.99) but not in TIMP-1A, WMWodds 0.99 (95% CI, 0.78–1.26). MMP-8 performed poorly in predicting 90-day mortality in both groups (AUC 0.63 [95% CI, 0.45–0.81], [ns] for elective patients and AUC 0.52 [95% CI, 0.46–0.58] for emergency patients). By contrast, TIMP-1A showed some predictive ability in the elective patient group (AUC 0.79 [95% CI, 0.65–0.92]) but little in the emergency group (AUC 0.62 [95% CI, 0.56–0.68]) in predicting 90-day mortality.

Back to Top | Article Outline


We demonstrated that serum levels of TIMP-1 and possibly MMP-8 are associated with increased mortality in a large unselected group of patients treated with mechanical ventilation for >6 hours. Admission TIMP-1 was found to be an independent predictor of mortality, although TIMP-1 performed better in the ARDS subgroup than in the whole cohort in predicting 90-day mortality.

To the best of our knowledge, this is the first study to show an association of TIMP-1 with increased mortality in a large population of critically ill patients with ARF. We and others have reported an association of elevated TIMP-1 with increased mortality in septic patients.11–13 In 1 study of 192 septic patients, elevated TIMP-1 and lower MMP-9 were associated with increased mortality.13 The AUC values for TIMP-1 were comparable with our findings. However, the optimal cutoff value was slightly higher (531 ng/mL) than the 458.6 ng/mL in our study.13 The most likely explanation is the uncertainty (wide 95% CI) of the cutoff in our study. In a smaller study of patients with severe sepsis, plasma levels of TIMP-1 performed well in discriminating survivors from nonsurvivors: AUC for TIMP-1 was 0.78, and patients with TIMP-1 levels >3200 ng/mL had a 4.5 times higher risk of dying than others.11 The cutoff level was higher than in our study, which may reflect differences in patient populations. In septic patients, systemic TIMP-1 levels may be higher generally than in our more heterogeneous patient population. Of our patients, only 13% (N = 68) had sepsis within 48 hours of admission. In our subgroup analysis, patients with sepsis had higher MMP-8A and TIMP-1A levels than nonseptic patients, but TIMP-1A did not differ between 90-day survivors and 90-day nonsurvivors. This may have been due to the relatively low number of patients with sepsis in the cohort. Notably, the ranges of the TIMP-1A levels in surviving and nonsurviving patients overlapped. Thus, TIMP-1 on its own may not be a reliable predictor of survival.

In the ARDS subgroup, TIMP-1 was elevated, and high levels were associated with increased mortality. In ARDS patients, an association of TIMP-1 with mortality has been reported.21 Decreased MMP-9/TIMP-1 ratio in bronchoalveolar lavage fluid has been associated with prolonged ARDS and worse prognosis.21,22 In these studies, the main focus was on MMPs, and elevated TIMP-1 had been interpreted in the context of MMP inhibition. TIMP-1 inhibits both MMP-8 and MMP-9 by noncovalently binding to the active part of MMP. In normal physiology, this balance between proteinase and inhibitor is strictly regulated, and an excess of uninhibited proteinase may cause tissue damage.8 In addition to the MMP-inhibiting activity, other biological functions of TIMP-1 have been described. TIMP-1 is reported to promote fibroblast and other cell proliferation, and it possesses antiapoptotic and proinflammatory effects.23–26 TIMP-1 is expressed in a variety of cells that participate in acute inflammatory reactions.23,24 Expression of TIMP-1 is stimulated by various substances, including platelet-derived growth factors, basic fibroblast growth factor, epidermal growth factor, and interleukins −1, −6, and −10,23–27 and it is elevated in hyperoxia-induced lung injury.28,29 Platelet-derived growth factors and basic fibroblast growth factor have a profibrotic effect in acute lung injury.30,31

Interestingly, in our study, TIMP-1 levels were associated with severity of oxygenation disturbance. This may be explained by an increased oxidative stress in the groups of worse gas exchange. TIMP-1 has been shown to be elevated in connection with hyperoxia in the lung28,29,32 and in oxidative damage to the venous endothelium.33 Unfortunately, measuring the level of oxidative stress or a surrogate other than the PaO2/FIO2 ratio was not possible with our data. Another explanation may lie in the functions of TIMP-1 that are independent of MMP inhibitory activity.23–26 Hypothetically, the stimulation of fibrotic processes, simultaneous inhibition of proteolysis, and promotion of proinflammatory processes by TIMP-1 could worsen the outcome in ARDS. Also, the antiapoptotic effect could inhibit the clearance of inflammatory cells from the lung. The biological significance of TIMP-1 in ARDS is a subject for further studies.

In our study, MMP-8 levels may have differed between survivors and nonsurvivors. However, the values were largely overlapping, and MMP-8 predicted outcome poorly. This is in broad accordance with Fligiel et al.,34 who noted in a smaller study that MMP-8 and MMP-9 were not associated with disease severity or mortality of ARDS patients. To the best of our knowledge, there are no previous studies on an association of systemic MMP-8 with mortality in a large cohort of ARF patients. MMP-8 and MMP-9 have been investigated in diverse diagnostic groups, and they tend to be elevated in acute illness10–12,21,22,34–37 and chronic lung diseases.38–40

MMPs play a role in many stages of acute inflammatory reactions. Their synthesis is induced, and their degranulation from leukocytes is stimulated by proinflammatory mediators. Local production of the enzyme has been demonstrated as well.7 Preclinical studies provide evidence of both harmful and beneficial roles in lung injury.14 In a clinical study of pediatric patients, elevated MMP-8 in tracheal aspirate samples was associated with increased duration of mechanical ventilation, but the patient population was too small to evaluate mortality.10

Fligiel et al.34 analyzed different MMPs in the bronchoalveolar lavage fluids of 28 ARDS patients. They found elevated levels of neutrophil-derived MMP-8 and MMP-9 and TIMP-1, and their strong correlation with neutrophil count was seen. Similarly, in our study, MMP-8 and TIMP-1 levels correlated with CRP and WBC count. MMP-8 levels were similar in the ARDS subgroup to those in the whole cohort. MMP-8 performed poorly as a predictive marker also in this subgroup. According to our results, systemic MMP-8 is not useful as a biomarker in ARF or ARDS.

There are several explanations for why MMP-8 and TIMP-1 performed differently in predicting mortality, although TIMP-1 is a MMP-8 inhibitor. First, systemic MMP-8 may not reflect the activity or concentration of the enzyme at the tissue level. In most studies, MMP-8 has been measured in bronchoalveolar lavage and tracheal aspirate samples, and it is possible that systemic levels reflect local processes poorly. MMP-8 levels were considerably higher in peritoneal fluid than in serum in a study of patients with peritonitis, and the levels did not correlate.37 MMP-8 levels are likely highest in the compartment where neutrophils accumulate because neutrophils are the principal source of MMP-8. TIMP-1, in turn, is released by several cell types, with the exception of neutrophils, and is found in most body fluids.24

The regulation of MMP-8 and TIMP-1 is not identical. First, depending on cell type, some cytokines may either upregulate or downregulate TIMP-1.41,42 Second, in serum, MMP-8 is inhibited by nonspecific plasma antiproteinases, for example, α2-macroglobulin.43 Methods used for measuring MMP-8, such as ELISA and IFMA, do not differentiate between the enzyme bound to its inhibitor and the free active enzyme. Therefore, measuring systemic MMP-8 levels by these methods may not reflect the true amount of biologically active enzyme. MMP-8 may also avoid its inhibitors at the tissue level when bound to the cell membrane of activated neutrophils.44 Third, TIMP-1 may have several functions independent of MMP inhibition.23–26 Finally, MMP-8 may have a dual role in inflammation, possibly initially harmful but later important in healing processes.14 This would explain why expression of high levels of the enzyme on admission is not associated with long-term outcome.

Mechanical ventilation per se may affect MMP-845 and MMP-9.46 In two-thirds of our patients, MMP-8 levels increased from admission to 48 hours later (data not shown). Due to different courses in the baseline diseases, our data were insufficient to analyze the relationship between mechanical ventilation and MMP-8 levels.

In this large, prospective, multicenter study, we included a clinically relevant group of critically ill patients, which may be seen as a strength of the study. Because different diagnostic groups were included, our results are generalizable. Our findings do not have direct clinical implications. However, biomarker panels for predictive purposes have recently been tested in a clinical study.47 According to our study, TIMP-1 is an independent predictor of poor outcome and may thus have potential in clinical outcome prediction as a component of a panel of selected biomarkers. This is a subject for future research.

Some limitations of the study must be addressed. First, we included patients who received invasive and noninvasive mechanical ventilation. The severity of disease may have been different between the groups. However, most patients received invasive ventilation, and this was the only group in which TIMP-1 levels differed between survivors and nonsurvivors. Our results may be generalizable for invasively ventilated patients, but they should be interpreted with caution for noninvasively ventilated patients. The large unselected group of patients also include patients not fitting the description “critically ill” despite the need for prolonged mechanical ventilation. Inclusion of elective postoperative patients may have biased our results. However, in a subgroup analysis of elective and emergency patients, the results were similar to those for the entire cohort. TIMP-1 had some predictive ability in discriminating 90-day survivors from nonsurvivors in patients needing mechanical ventilation after an elective operation. A relatively large proportion of cardiac surgical patients in the electively operated patient group may explain why TIMP-1 also had some predictive ability in this group. TIMP-1 has been shown to predict adverse cardiac events in patients with cardiac disease.48,49

The second limitation is that our data lacked information of admission lactate levels, which are important in outcome prediction.50 Furthermore, an extended analysis of cytokine levels and MMP-9 would have facilitated interpretation of the inflammatory status of different patient groups. In the subgroup, serial measurements of our markers and other inflammatory variables could have enhanced our knowledge about the time course of these markers throughout the disease.

The last limitation is the study size. According to a post hoc sample size calculation, based on 95% CIs for AUC as in a previous study,51 and the number of 90-day deaths in this study population, a sample size of at least 500 patients would yield clinically precise AUCs with 95% CIs <0.15 with 90-day mortality. The sample size calculation was only an approximate estimation due to an unknown distribution of measured laboratory variables. Nevertheless, the current study is one of the largest in this field.

In conclusion, high levels of TIMP-1 were independently associated with increased 90-day mortality in a large group of critically ill, mechanically ventilated patients. However, serum MMP-8 levels did not predict outcome in ARF or ARDS patients.

Back to Top | Article Outline

Appendix. Participants in the FINNALI Study

Central Finland Central Hospital: Dr. Raili Laru-Sompa, Tiina Kirkhope, Sirpa Nykänen; East Savo Central Hospital: Dr. Markku Suvela, Sirpa Kauppinen, Anne-Marja Turkulainen; Helsinki University Central Hospital (HUCH)-Jorvi Hospital: Dr. Tero Varpula, Mira Rahkonen; HUCH-Meilahti Hospital ICU: Dr. Anne Kuitunen, Dr. Anna-Maija Korhonen, Dr. Rita Linko, Dr. Marjatta Okkonen, Janne Myller, Jarmo Pekkola, Leena Pettilä, Sari Sutinen; HUCH-Meilahti Hospital, Cardiac Surgical ICU: Dr. Raili Suojaranta-Ylinen, Dr. Sinikka Kukkonen, Elina Nurmi-Toivonen; HUCH-Meilahti Hospital, Department of Medicine: Dr. Tom Bäcklund, Dr. Juhani Rossinen, Riina Mäkelä; HUCH-Töölö Hospital: Dr. Janne Reitala, Dr. Jyrki Vuola, Raija Niemi, Marja-Leena Pihlajamaa, Aira Uusipaavalniemi; HUCH-Surgical Hospital: Dr. Anna-Maria Koivusalo, Pasi Kyllönen; Kanta-Häme Central Hospital: Dr. Risto Puolakka, Tarja Heikkilä, Piia Laitinen; Keski-Pohjanmaa Central Hospital: Dr. Samuli Forsström, Dr. Tadeusz Kaminski, Tuija Kuusela; Kuopio University Hospital: Dr. Esko Ruokonen, Dr. Ilkka Parviainen, Elina Halonen, Sari Rahikainen; Kymenlaakso Central Hospital: Dr. Jussi Pentti, Dr. Seija Alila, Reija Koskinen; Lappi Central Hospital: Dr. Outi Kiviniemi, Tarja Laurila, Tiina Pikkuhookana; North Karelia Central Hospital: Dr. Matti Reinikainen, Tanja Eiserbeck, Helena Jyrkönen; Oulu University Hospital: Dr. Tero Ala-Kokko, Tarja Lamberg, Sinikka Sälkiö; Päijät-Häme Central Hospital: Dr. Pekka Loisa; Satakunta Central Hospital: Dr. Vesa Lund, Pauliina Perkola, Päivi Tuominen; Seinäjoki Central Hospital: Dr. Kari Saarinen, Dr. Matti Viitanen, Niina Siirilä, Johanna Soini; South Karelia Central Hospital: Dr. Seppo Hovilehto, Dr. Anne Kirsi, Dr. Pekka Tiainen, Sanna Asikainen; South Savo Central Hospital: Dr. Heikki Laine, Pekka Kettunen, Kirsi Reponen; Tampere University Hospital: Dr. Sari Karlsson, Dr. Jyrki Tenhunen, Anna-Liina Korkala, Samuli Kortelainen, Sanna Mäkinen, Minna-Liisa Peltola; Turku University Hospital: Dr. Juha Perttilä, Dr. Olli Arola, Dr. Outi Inkinen, Dr. Erkki Kentala, Jutta Kotamäki; Vaasa Central Hospital: Dr. Pentti Kairi.

Back to Top | Article Outline


Name: Johanna Hästbacka, MD, PhD.

Contribution: This author contributed to the study concept, conducted the statistical analysis and interpretation of the data, and prepared the manuscript.

Attestation: Johanna Hästbacka has reviewed the original study data and the data analysis. Johanna Hästbacka is the archival author.

Name: Rita Linko, MD, PhD.

Contribution: This author helped contributing to the study concept, collecting the clinical data, and drafting the manuscript.

Attestation: Rita Linko has approved the final manuscript.

Name: Taina Tervahartiala, DDS, PhD.

Contribution: This author contributed to the laboratory analysis of the samples and drafting the manuscript.

Attestation: Taina Tervahartiala has approved the final manuscript.

Name: Tero Varpula, MD, PhD.

Contribution: This author helped contributing to the study concept, collecting the data, and drafting the manuscript.

Attestation: Tero Varpula has approved the final manuscript.

Name: Seppo Hovilehto, MD.

Contribution: This author contributed to collecting the clinical data and laboratory samples and drafting the manuscript.

Attestation: Seppo Hovilehto has approved the final manuscript.

Name: Ilkka Parviainen, MD, PhD.

Contribution: This author contributed to collecting the clinical data and laboratory samples and drafting the manuscript.

Attestation: Ilkka Parviainen has approved the final manuscript.

Name: Suvi T Vaara, MD, PhD.

Contribution: This author contributed to analyzing the results and drafting the manuscript.

Attestation: Suvi Vaara has approved the final manuscript.

Name: Timo Sorsa, DDS, PhD.

Contribution: This author contributed to the analysis of the laboratory samples and drafting the manuscript.

Attestation: Timo Sorsa has approved the final manuscript.

Name: Ville Pettilä, MD, PhD.

Contribution: This author helped contributing to the study concept, collecting the clinical data, analyzing the results, and drafting the manuscript.

Attestation: Ville Pettilä attests to having reviewed the original study data and data analysis and having approved the final manuscript.

This manuscript was handled by: Steven L. Shafer, MD.

Back to Top | Article Outline


We thank Carol Pelli for expert language editing and Professor Antti Nevanlinna for statistical advice.

Back to Top | Article Outline


1. Vincent JL, Akca S, De Mendonca A, Haji-Michael P, Sprung C, Moreno R, Antonelli M, Suter PM. The epidemiology of acute respiratory failure in critically ill patients. Chest. 2002;121:1602–9

2. Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND, Caldwell E, Fan E, Camporota L, Slutsky AS. Acute respiratory distress syndrome: the Berlin definition. JAMA. 2012;307:2526–33

3. Grommes J, Soehnlein O. Contribution of neutrophils to acute lung injury. Mol Med. 2011;17:293–307

4. Owen CA, Campbell EJ. The cell biology of leukocyte-mediated proteolysis. J Leukoc Biol. 1999;65:137–50

5. Sternlicht MD, Werb Z. How matrix metalloproteinases regulate cell behavior. Annu Rev Cell Dev Biol. 2001;17:463–516

6. Van Lint P, Libert C. Matrix metalloproteinase-8: cleavage can be decisive. Cytokine Growth Factor Rev. 2006;17:217–23

7. Vanlaere I, Libert C. Matrix metalloproteinases as drug targets in infections caused by gram-negative bacteria and in septic shock. Clin Microbiol Rev. 2009;22:224–39

8. Visse R, Nagase H. Matrix metalloproteinases and tissue inhibitors of metalloproteinases: structure, function, and biochemistry. Circ Res. 2003;92:827–39

9. Elkington PT, O’Kane CM, Friedland JS. The paradox of matrix metalloproteinases in infectious disease. Clin Exp Immunol. 2005;142:12–20

10. Kong MY, Li Y, Oster R, Gaggar A, Clancy JP. Early elevation of matrix metalloproteinase-8 and -9 in pediatric ARDS is associated with an increased risk of prolonged mechanical ventilation. PLoS One. 2011;6:e22596

11. Hoffmann U, Bertsch T, Dvortsak E, Liebetrau C, Lang S, Liebe V, Huhle G, Borggrefe M, Brueckmann M. Matrix-metalloproteinases and their inhibitors are elevated in severe sepsis: prognostic value of TIMP-1 in severe sepsis. Scand J Infect Dis. 2006;38:867–72

12. Lauhio A, Hästbacka J, Pettilä V, Tervahartiala T, Karlsson S, Varpula T, Varpula M, Ruokonen E, Sorsa T, Kolho E. Serum MMP-8, -9 and TIMP-1 in sepsis: high serum levels of MMP-8 and TIMP-1 are associated with fatal outcome in a multicentre, prospective cohort study. Hypothetical impact of tetracyclines. Pharmacol Res. 2011;64:590–4

13. Lorente L, Martín MM, Labarta L, Díaz C, Solé-Violán J, Blanquer J, Orbe J, Rodríguez JA, Jiménez A, Borreguero-León JM, Belmonte F, Medina JC, Llimiñana MC, Ferrer-Agüero JM, Ferreres J, Mora ML, Lubillo S, Sánchez M, Barrios Y, Sierra A, Páramo JA. Matrix metalloproteinase-9, -10, and tissue inhibitor of matrix metalloproteinases-1 blood levels as biomarkers of severity and mortality in sepsis. Crit Care. 2009;13:R158

14. Davey A, McAuley DF, O’Kane CM. Matrix metalloproteinases in acute lung injury: mediators of injury and drivers of repair. Eur Respir J. 2011;38:959–70

15. Linko R, Okkonen M, Pettilä V, Perttilä J, Parviainen I, Ruokonen E, Tenhunen J, Ala-Kokko T, Varpula T. Acute respiratory failure in intensive care units. FINNALI: A prospective cohort study. Intensive Care Med. 2009;35:1352–61

16. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L, Lamy M, Legall JR, Morris A, Spragg R. The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med. 1994;149:818–24

17. Hanemaaijer R, Sorsa T, Konttinen YT, Ding Y, Sutinen M, Visser H, van Hinsbergh VW, Helaakoski T, Kainulainen T, Rönkä H, Tschesche H, Salo T. Matrix metalloproteinase-8 is expressed in rheumatoid synovial fibroblasts and endothelial cells. Regulation by tumor necrosis factor-alpha and doxycycline. J Biol Chem. 1997;272:31504–9

18. Tuomainen AM, Nyyssönen K, Laukkanen JA, Tervahartiala T, Tuomainen TP, Salonen JT, Sorsa T, Pussinen PJ. Serum matrix metalloproteinase-8 concentrations are associated with cardiovascular outcome in men. Arterioscler Thromb Vasc Biol. 2007;27:2722–8

19. Divine G, Norton HJ, Hunt R, Dienemann J. A review of analysis and sample size calculation considerations for Wilcoxon tests. Anesth Analg. 2013;117:699–710

20. Youden WJ. Index for rating diagnostic tests. Cancer. 1950;3:32–5

21. Ricou B, Nicod L, Lacraz S, Welgus HG, Suter PM, Dayer JM. Matrix metalloproteinases and TIMP in acute respiratory distress syndrome. Am J Respir Crit Care Med. 1996;154:346–52

22. Lanchou J, Corbel M, Tanguy M, Germain N, Boichot E, Theret N, Clement B, Lagente V, Malledant Y. Imbalance between matrix metalloproteinases (MMP-9 and MMP-2) and tissue inhibitors of metalloproteinases (TIMP-1 and TIMP-2) in acute respiratory distress syndrome patients. Crit Care Med. 2003;31:536–42

23. Murate T, Hayakawa T. Multiple functions of tissue inhibitors of metalloproteinases (TIMPs): new aspects in hematopoiesis. Platelets. 1999;10:5–16

24. Lambert E, Dassé E, Haye B, Petitfrère E. TIMPs as multifacial proteins. Crit Rev Oncol Hematol. 2004;49:187–98

25. Guedez L, Stetler-Stevenson WG, Wolff L, Wang J, Fukushima P, Mansoor A, Stetler-Stevenson M. In vitro suppression of programmed cell death of B cells by tissue inhibitor of metalloproteinases-1. J Clin Invest. 1998;102:2002–10

26. Apparailly F, Noël D, Millet V, Baker AH, Lisignoli G, Jacquet C, Kaiser MJ, Sany J, Jorgensen C. Paradoxical effects of tissue inhibitor of metalloproteinases 1 gene transfer in collagen-induced arthritis. Arthritis Rheum. 2001;44:1444–54

27. Mostafa Mtairag E, Chollet-Martin S, Oudghiri M, Laquay N, Jacob MP, Michel JB, Feldman LJ. Effects of interleukin-10 on monocyte/endothelial cell adhesion and MMP-9/TIMP-1 secretion. Cardiovasc Res. 2001;49:882–90

28. Ward NS, Waxman AB, Homer RJ, Mantell LL, Einarsson O, Du Y, Elias JA. Interleukin-6-induced protection in hyperoxic acute lung injury. Am J Respir Cell Mol Biol. 2000;22:535–42

29. Veness-Meehan KA, Cheng ER, Mercier CE, Blixt SL, Johnston CJ, Watkins RH, Horowitz S. Cell-specific alterations in expression of hyperoxia-induced mRNAs of lung. Am J Respir Cell Mol Biol. 1991;5:516–21

30. Henke C, Marineili W, Jessurun J, Fox J, Harms D, Peterson M, Chiang L, Doran P. Macrophage production of basic fibroblast growth factor in the fibroproliferative disorder of alveolar fibrosis after lung injury. Am J Pathol. 1993;143:1189–99

31. Walsh J, Absher M, Kelley J. Variable expression of platelet-derived growth factor family proteins in acute lung injury. Am J Respir Cell Mol Biol. 1993;9:637–44

32. Pardo A, Barrios R, Maldonado V, Meléndez J, Pérez J, Ruiz V, Segura-Valdez L, Sznajder JI, Selman M. Gelatinases A and B are up-regulated in rat lungs by subacute hyperoxia: pathogenetic implications. Am J Pathol. 1998;153:833–44

33. Moore R, Hawley A, Sigler R, Farris D, Wrobleski S, Ramacciotti E, Myers D. Tissue inhibitor of metalloproteinase-1 is an early marker of acute endothelial dysfunction in a rodent model of venous oxidative injury. Ann Vasc Surg. 2009;23:498–505

34. Fligiel SE, Standiford T, Fligiel HM, Tashkin D, Strieter RM, Warner RL, Johnson KJ, Varani J. Matrix metalloproteinases and matrix metalloproteinase inhibitors in acute lung injury. Hum Pathol. 2006;37:422–30

35. Hartog CM, Wermelt JA, Sommerfeld CO, Eichler W, Dalhoff K, Braun J. Pulmonary matrix metalloproteinase excess in hospital-acquired pneumonia. Am J Respir Crit Care Med. 2003;167:593–8

36. O’Kane CM, McKeown SW, Perkins GD, Bassford CR, Gao F, Thickett DR, McAuley DF. Salbutamol up-regulates matrix metalloproteinase-9 in the alveolar space in the acute respiratory distress syndrome. Crit Care Med. 2009;37:2242–9

37. Hästbacka J, Hynninen M, Kolho E, Pettilä V, Tervahartiala T, Sorsa T, Lauhio A. Collagenase 2/matrix metalloproteinase 8 in critically ill patients with secondary peritonitis. Shock. 2007;27:145–50

38. Ratjen F, Hartog CM, Paul K, Wermelt J, Braun J. Matrix metalloproteases in BAL fluid of patients with cystic fibrosis and their modulation by treatment with dornase alpha. Thorax. 2002;57:930–4

39. Segura-Valdez L, Pardo A, Gaxiola M, Uhal BD, Becerril C, Selman M. Upregulation of gelatinases A and B, collagenases 1 and 2, and increased parenchymal cell death in COPD. Chest. 2000;117:684–94

40. Sepper R, Konttinen YT, Ding Y, Takagi M, Sorsa T. Human neutrophil collagenase (MMP-8), identified in bronchiectasis BAL fluid, correlates with severity of disease. Chest. 1995;107:1641–7

41. Yao PM, Maitre B, Delacourt C, Buhler JM, Harf A, Lafuma C. Divergent regulation of 92-kDa gelatinase and TIMP-1 by HBECs in response to IL-1beta and TNF-alpha. Am J Physiol. 1997;273:L866–74

42. Overall CM. Regulation of tissue inhibitor of matrix metalloproteinase expression. Ann N Y Acad Sci. 1994;732:51–64

43. Sottrup-Jensen L, Birkedal-Hansen H. Human fibroblast collagenase-alpha-macroglobulin interactions. Localization of cleavage sites in the bait regions of five mammalian alpha-macroglobulins. J Biol Chem. 1989;264:393–401

44. Owen CA, Hu Z, Lopez-Otin C, Shapiro SD. Membrane-bound matrix metalloproteinase-8 on activated polymorphonuclear cells is a potent, tissue inhibitor of metalloproteinase-resistant collagenase and serpinase. J Immunol. 2004;172:7791–803

45. Albaiceta GM, Gutierrez-Fernández A, García-Prieto E, Puente XS, Parra D, Astudillo A, Campestre C, Cabrera S, Gonzalez-Lopez A, Fueyo A, Taboada F, López-Otin C. Absence or inhibition of matrix metalloproteinase-8 decreases ventilator-induced lung injury. Am J Respir Cell Mol Biol. 2010;43:555–63

46. Albaiceta GM, Gutiérrez-Fernández A, Parra D, Astudillo A, García-Prieto E, Taboada F, Fueyo A. Lack of matrix metalloproteinase-9 worsens ventilator-induced lung injury. Am J Physiol Lung Cell Mol Physiol. 2008;294:L535–43

47. Calfee C, Ware L, Glidden D, Eisner M, Parsons P, Thompson B, Matthay M. National heart, blood, and lung institute acute respiratory distress network: Use of risk reclassification with multiple biomarkers improves mortality prediction in acute lung injury. Crit Care Med. 2011;39:711–7

48. Kelly D, Squire IB, Khan SQ, Dhillon O, Narayan H, Ng KH, Quinn P, Davies JE, Ng LL. Usefulness of plasma tissue inhibitors of metalloproteinases as markers of prognosis after acute myocardial infarction. Am J Cardiol. 2010;106:477–82

49. Cavusoglu E, Ruwende C, Chopra V, Yanamadala S, Eng C, Clark LT, Pinsky DJ, Marmur JD. Tissue inhibitor of metalloproteinase-1 (TIMP-1) is an independent predictor of all-cause mortality, cardiac mortality, and myocardial infarction. Am Heart J. 2006;151:1101.e1–8

50. Nichol A, Bailey M, Egi M, Pettila V, French C, Stachowski E, Reade MC, Cooper DJ, Bellomo R. Dynamic lactate indices as predictors of outcome in critically ill patients. Crit Care. 2011;15:R242

51. Haase M, Bellomo R, Devarajan P, Ma Q, Bennett MR, Möckel M, Matalanis G, Dragun D, Haase-Fielitz A. Novel biomarkers early predict the severity of acute kidney injury after cardiac surgery in adults. Ann Thorac Surg. 2009;88:124–30

© 2014 International Anesthesia Research Society


Become a Society Member