Abstract
Objective: To evaluate the performance of cystatin C as a prognostic and predictive marker in a trial of patients with acute respiratory distress syndrome (ARDS).
Design, patients and setting: A retrospective analysis was performed on plasma samples from patients included in the HARP-2 (hydroxymethylglutaryl-CoA reductase inhibition with simvastatin in acute lung injury to reduce pulmonary dysfunction) trial — a multicentre, phase 2b trial carried out in general intensive care units across 40 hospitals in the United Kingdom and Ireland. Cystatin C concentrations in plasma obtained from 466 patients with ARDS (before they were randomly assigned in the trial) were quantified by ELISA (enzyme-linked immunosorbent assay).
Results: In a univariate analysis, plasma cystatin C concentrations were significantly higher in patients with ARDS who did not survive past 28 days (odds ratio [OR], 1.39 [95% CI, 1.12–1.72]; P = 0.002). In a multivariate model adjusted for selected covariates, cystatin C concentrations remained higher among patients with ARDS who did not survive, although this did not reach statistical significance (OR, 1.28 [95% CI, 0.96–1.71]; P = 0.090). Cystatin C concentration was also significantly associated with hyperinflammatory ARDS (OR, 2.64 [95% CI, 1.83–3.89]; P < 0.001). In multivariate models adjusted for both cystatin C concentration and ARDS subphenotype, hyperinflammatory ARDS was prognostic for mortality (OR, 2.06 [95% CI, 1.16–3.64]; P = 0.013) but cystatin C concentration was not (OR, 1.16 [95% CI, 0.85–1.57]; P = 0.346). In a multivariate analysis, hyperinflammatory ARDS was predictive of a beneficial effect of simvastatin on mortality (OR, 2.05 [95% CI, 1.16–3.62]; P = 0.014) but cystatin C concentration was not (OR, 1.10 [95% CI, 0.77–1.56]; P = 0.614).
Conclusion: The association between cystatin C concentration and mortality in ARDS may be dependent on inflammatory subphenotype. Cystatin C concentration does not appear to add to existing prognostic or predictive approaches.
Cystatin C is a small, basic protein, which is regarded as the canonical extracellular inhibitor of the cysteine protease family of enzymes. During homeostasis, cystatin C is thought to be produced at a constant rate by all nucleated cells.1, 2 This fact — along with its free filtration at the glomerulus, its complete reabsorption and catabolism by the proximal tubules, and the lack of cystatin C secretion by the renal tubules — has led to serum cystatin C being regarded as an alternative to serum creatinine for assessing renal function.3, 4 Recent work has suggested that cystatin C may have prognostic and predictive utility in conditions beyond kidney injury. In the context of critical illness, serum cystatin C within the general critical care population has been correlated with mortality in patients with and without kidney injury as defined using conventional creatinine-based methods.5 Furthermore, baseline circulating cystatin C levels were shown to be associated with mortality in patients with acute respiratory distress syndrome (ARDS) in a recent substudy of the Fluid and Catheter Treatment Trial (FACTT).6 In a study of critically ill patients with coronavirus disease 2019 (COVID-19), patients with higher cystatin C levels had a higher incidence of ARDS, as well as increased inflammatory indices and Acute Physiology and Chronic Health Evaluation (APACHE) II scores.7 Although these studies suggest that measuring cystatin C has prognostic value in these patients, the findings require verification in additional cohorts of critically ill patients and those with ARDS to determine whether they are generalisable.
In several trials, latent class analysis (LCA) has identified two subphenotypes of ARDS that have different mortality rates and different responses to therapy.8, 9, 10 Interestingly, it has been reported that plasma cystatin C levels are higher among patients in the hyperinflammatory subphenotype, which is characterised by high concentrations of inflammatory biomarkers, shock, metabolic acidosis and worse clinical outcomes.6, 8 Thus, cystatin C may also have some potential as a biomarker for ARDS subphenotypes. However, to our knowledge, this relationship has not been observed in any other studies.
In the present study, we aimed to explore the prognostic value of plasma cystatin C levels in a cohort of patients with ARDS recruited for the HARP-2 (hydroxymethylglutaryl-CoA reductase inhibition with simvastatin in acute lung injury to reduce pulmonary dysfunction) trial. We also sought to explore whether cystatin C levels could predict a positive response to simvastatin therapy in these patients.
Methods
The HARP-2 trial was a multicentre, phase 2b randomised controlled trial of simvastatin versus placebo carried out in general intensive care units across 40 hospitals in the United Kingdom and Ireland.11 Briefly, adult patients were included in the trial if they were intubated, mechanically ventilated and were within 48 hours after the onset of ARDS. ARDS was defined by the presence of bilateral pulmonary infiltrates on chest radiography, no evidence of left atrial hypertension, and a ratio of the partial pressure of arterial oxygen (Pao2) to the fraction of inspired oxygen (Fio2) of ≤ 300 mmHg (≤ 40 kPa). Patients with severe renal impairment who were not receiving renal replacement therapy were excluded from the trial. Other exclusion criteria are detailed in the study protocol.12 Patient characteristics were recorded at the time of admission. At enrolment, patients were randomised in a 1:1 ratio to once-daily enteral simvastatin (80 mg) or placebo, which they received for up to 28 days. The primary outcome measure was ventilator-free days to 28 days after randomisation. The full list of secondary outcome measures can be found in the primary publication and the accompanying study protocol.11, 12 Although simvastatin was shown to be safe, there were no significant differences in clinical outcomes between the two study groups.
Stratification of patients
As part of a previous analysis of the HARP-2 trial, patients were divided into those with a hypoinflammatory subphenotype and those with a hyperinflammatory subphenotype, as identified by LCA.9 For the present study, in the absence of measured glomerular filtration rate data, we also stratified patients based on baseline creatinine concentration, either < 170 μmol/L (deemed to be within the normal range) or ≥ 170 μmol/L.
Cystatin C measurement
Cystatin C in plasma was quantified using the Human Cystatin C DuoSet ELISA (enzyme-linked immunosorbent assay) (R&D Systems). During the HARP-2 trial, plasma was generated from heparinised blood drawn before randomisation and was stored at –70°C until required for analysis. Assays were performed by a single researcher, avoiding any inter-user or inter-laboratory variability in measurement. Samples were diluted to the appropriate dilution in 1% bovine serum albumin (Sigma-Aldrich) in phosphate-buffered saline and assayed as per the manufacturer’s instructions. The threshold of detection for the assay was 62.5 pg/mL.
Statistical analysis
Summary statistics for all relevant prognostic factors by treatment group and overall are presented. To account for the skewness in the distribution of the data (Online Appendix, figure 1), cystatin C concentrations were transformed using the natural logarithm (ln) and standardised by subtraction of the mean and division by the standard deviation of the mean to improve numerical stability. Subsequent analyses were performed using these ln-transformed and standardised values. Serum creatinine concentrations were scaled to give a mean of 0 and variance of 1 for ease of interpretation alongside cystatin C concentrations. Boxplots were produced to illustrate the relationships between mortality, LCA subphenotype and baseline cystatin C concentration. The relationships between baseline cystatin C concentrations and binary outcomes (mortality and subphenotype) were explored by fitting logistic regression models, showing both univariate and adjusted (multivariate) estimates. Pre-defined clinically relevant variables (age, treatment group, vasopressor usage, Pao2:Fio2 ratio and serum creatinine concentration) were included in the models. The datasets that we analysed are available from the corresponding author on reasonable request.
Ethics approval
The original HARP-2 trial was approved by the Office for Research Ethics Committees Northern Ireland (10/NIR02/36) and by the institutional research ethics committee for each study site. The trial is registered on the ISRCTN Registry (ISRCTN88244364).
Results
The baseline characteristics of the 466 patients included in this analysis of the HARP-2 trial are summarised in Table 1. The mean age of patients was 54 years and 56.7% of patients were men. The overall 28-day mortality rate for this cohort was 24.9%. The most common primary causes of ARDS were pneumonia and sepsis. Mean APACHE II and Sequential Organ Failure Assessment scores were 18.9 and 8.6, respectively, and 11.7% of patients had a baseline creatinine concentration of ≥ 170 μmol/L. The median concentration of cystatin C in plasma at baseline was 2.03 μg/mL and the interquartile range was 1.45–2.94 μg/mL.
Table 1.
Baseline characteristics of patients from HARP-2 trial included in analysis⁎
| Simvastatin (n = 226) | Placebo (n = 240) | Total (n = 466) | |
|---|---|---|---|
| Age, years | 52.8 ± 16.1 | 55.1 ± 16.6 | 54.0 ± 16.4 |
| Men | 121 (53.5%) | 143 (59.6%) | 264 (56.7%) |
| Sepsis | 166 (73.5%) | 184 (76.7%) | 350 (75.1 %) |
| Cause of ARDS† | |||
| Smoke or toxin inhalation | 0 | 2 (0.8%) | 2 (0.4%) |
| Gastric content aspiration | 18 (8.0%) | 24 (10.0%) | 42 (9.0%) |
| Thoracic trauma | 20 (8.8%) | 9 (3.8%) | 29 (6.2%) |
| Pneumonia | 144 (63.7%) | 132 (55.0%) | 276 (59.2%) |
| Sepsis | 92 (40.7%) | 100 (41.7%) | 192 (41.2%) |
| Pancreatitis | 3 (1.3%) | 14 (5.8%) | 17 (3.6%) |
| Non-thoracic trauma | 4 (1.8%) | 7 (2.9%) | 11 (2.4%) |
| Other | 24 (10.6%) | 33(13.8%) | 57 (12.2%) |
| APACHE II score‡ | 19.3 ± 6.7 (n = 195) | 18.4 ± 6.4 (n = 215) | 18.9 ± 6.6 (n = 410) |
| SOFA score§ | 8.3 ± 3.2 (n = 192) | 8.8 ± 2.9 (n = 213) | 8.6 ± 3.0 (n = 405) |
| Serum creatinine concentration, ≥ 170 μmol/L | 23 (10.8%) | 29 (12.5%) | 52 (11.7%) |
| Serum creatinine concentration, μmol/L | 98.2 ± 63.3 (n = 214) | 108.9 ± 82.2 (n = 232) | 103.7 ± 73.9 (n = 446) |
| Vasopressor dependent | 142 (62.8%) | 155 (64.6%) | 297 (63.7%) |
| Lowest mean arterial pressure, mmHg | 65.7 ± 9.3 (n = 225) | 65.0 ± 8.6 (n = 239) | 65.3 ± 9.0 (n = 464) |
| Inspiratory plateau pressure, cmH2O | 23.6 ± 5.9 (n = 120) | 23.7 ± 6.0 (n = 134) | 23.7 ± 6.0 (n = 254) |
| Tidal volume, mL/kg of predicted bodyweight¶ | 8.1 ± 2.8 (n = 209) | 8.1 ± 2.6 (n = 221) | 8.1 ± 2.7 (n = 430) |
| Pao2:Fio2 ratio | |||
| mmHg | 123.0 ± 53.1 | 132.5 ± 54.7 | 127.9 ± 54.1 |
| kPa | 16.4 ± 7.1 | 17.7 ± 7.3 | 17.1 ± 7.2 |
| Oxygenation index, cmH2O/mmHg | 14.4 ± 10.4 (n = 169) | 14.6 ± 12.0 (n = 175) | 14.5 ± 11.2 (n = 344) |
| Alanine aminotransferase concentration, U/L | 42.7 ± 43.7 (n = 225) | 46.0 ± 44.1 (n = 239) | 44.4 ± 43.9 (n = 464) |
| Aspartate aminotransferase concentration, U/L | 57.2 ± 44.4 (n = 218) | 61.4 ± 60.4 (n = 227) | 59.3 ± 53.1 (n = 445) |
| Creatine kinase concentration, U/L | 329.4 ± 511.3 (n = 225) | 286.0 ± 480.8 (n = 237) | 307.1 ± 495.8 (n = 462) |
APACHE = Acute Physiology and Chronic Health Evaluation; ARDS = acute respiratory distress syndrome; Fio2 = fraction of inspired oxygen; Pao2 = partial pressure of arterial oxygen; SOFA = Sequential Organ Failure Assessment.
Data are mean ± standard deviation or number (percentage); denominators are shown in cases where data were missing for some patients.
Patients may have had more than one cause of ARDS identified.
Scores on the APACHE II scale range from 0 to 71, with higher scores indicating more severe disease.
Scores on the SOFA scale range from 0 to 24, with higher scores indicating more severe disease.
Predicted bodyweight was calculated as 2.3 kg for each inch of height above 60 in (152 cm) added to a base weight of 50.0 kg for men and 45.5 kg for women
For our initial analysis, patients were divided into survivors and non-survivors at 28 days after enrolment, those with creatinine concentrations ≥ 170 μmol/L and < 170 μmol/L, and those with hypoinflammatory and hyperinflammatory LCA subphenotype. The ln-transformed plasma cystatin C concentrations in these patient groups are presented in Figure 1. In a univariate analysis, cystatin C concentrations were higher in patients who did not survive (odds ratio [OR], 1.39 [95% CI, 1.12–1.72]; P = 0.002). When patients were stratified by creatinine concentration, cystatin C concentrations were higher in patients with higher creatinine concentrations. However, within the group with creatinine concentrations < 170 μmol/L, cystatin C concentrations were higher in those who did not survive (OR, 1.48 [1.16–1.89]; P = 0.002). The low number of patients with high serum creatinine concentrations (n = 52) restricted our ability to analyse this cohort, so the association between mortality and cystatin C concentration in these patients was not addressed.
Figure 1.
Box and whisker plots of natural logarithm-transformed baseline plasma cystatin C concentrations using different combinations of patient stratification*
When patients were stratified by LCA subphenotype, cystatin C concentrations were higher in patients with hyperinflammatory ARDS. Higher cystatin C concentrations were observed in non-survivors with hypoinflammatory ARDS (OR, 1.52 [95% CI, 1.07–2.17]; P = 0.021) but not in non-survivors with hyperinflammatory ARDS (OR, 0.81, [95% CI, 0.56–1.16]; P = 0.255).
To further assess the association between plasma cystatin C concentration and mortality, cystatin C was included in a multivariate prognostic model adjusted for pre-specified clinically important covariates (Table 2). In this multivariate model, the association between cystatin C concentration and mortality did not reach statistical significance (OR, 1.28 [95% CI, 0.96-1.71]; P = 0.090). Almost identical findings were reported when patients with high serum creatinine concentration (≥ 170 μmol/L) were excluded from the analysis (OR, 1.27 [95% CI, 0.98–1.65]; P = 0.076) (Online Appendix, table 1), indicating that any apparent association between cystatin C concentration and mortality was independent of serum creatinine concentration.
Table 2.
Adjusted associations between 28-day mortality, plasma cystatin C concentration and selected covariates
| Variate | Odds ratio (95% CI), univariable | P | Odds ratio (95% CI), multivariable | P |
|---|---|---|---|---|
| Age | 1.03 (1.02–1.04) | < 0.001 | 1.03 (1.01–1.05) | < 0.001 |
| Pao2:Fio2 ratio | 1.00 (0.99–1.00) | 0.368 | 1.00 (0.99–1.00) | 0.468 |
| Vasopressor used⁎ | 2.00 (1.26–3.25) | 0.004 | 1.84 (1.11–3.10) | 0.019 |
| Ln cystatin C† | 1.39 (1.12–1.72) | 0.002 | 1.28 (0.96–1.71) | 0.090 |
| Serum creatinine | 1.14 (0.93–1.40) | 0.198 | 0.89 (0.66–1.17) | 0.416 |
| Treated with simvastatin‡ | 0.68 (0.44–1.04) | 0.078 | 0.70 (0.44–1.10) | 0.124 |
Fio2 = fraction of inspired oxygen; Pao2 = partial pressure of arterial oxygen.
Reference category: vasopressor not used.
Natural logarithm-transformed baseline plasma cystatin C concentration.
Reference category: placebo.
Since ARDS subphenotype is strongly associated with mortality, and because an association between cystatin C concentration and ARDS subphenotype has previously been reported,6 we explored these relationships in the HARP-2 cohort. Cystatin C concentrations were significantly associated with hyperinflammatory ARDS in the univariate model (OR, 4.45 [95% CI, 3.34–6.09]; P < 0.001) and the multivariate model (OR, 2.64 [95% CI, 1.83–3.89]; P < 0.001) (Table 3). Also, serum creatinine was independently associated with hyperinflammatory ARDS in the multivariate model (OR, 3.04 [95% CI, 2.01–4.81]; P < 0.001).
Table 3.
Adjusted associations between latent class analysis subphenotype, plasma cystatin C concentration and selected covariates
| Variate | Odds ratio (95% CI), univariable | P | Odds ratio (95% CI), multivariable | P |
|---|---|---|---|---|
| Age | 1.04 (1.02–1.05) | < 0.001 | 1.02 (1.00–1.04) | 0.018 |
| Pao2:Fio2 ratio | 1.00 (0.99–1.00) | 0.052 | 1.00 (0.99–1.00) | 0.187 |
| Vasopressor used⁎ | 3.33 (2.15–5.28) | 0.001 | 3.98 (2.22–7.43) | < 0.001 |
| Ln cystatin C† | 4.45 (3.34–6.09) | < 0.001 | 2.64 (1.83–3.89) | < 0.001 |
| Serum creatinine | 6.26 (4.29–9.48) | < 0.001 | 3.04 (2.01–4.81) | < 0.001 |
| Treated with simvastatin‡ | 0.81 (0.55–1.19) | 0.279 | 0.86 (0.51–1.45) | 0.583 |
Fio2 = fraction of inspired oxygen; Pao2 = partial pressure of arterial oxygen.
Reference category: vasopressor not used.
Natural logarithm-transformed baseline plasma cystatin C concentration.
Reference category: placebo.
Interestingly, when ARDS subphenotype was added to the model described in Table 2, no association between cystatin C concentration and mortality was seen (OR, 1.16 [95% CI, 0.85–1.57]; P = 0.346). However, as expected, hyperinflammatory ARDS was significantly associated with mortality (OR, 2.06 [95% CI, 1.16–3.64]; P = 0.013) (Table 4). This finding suggested that the weak association between plasma cystatin C concentrations and 28-day mortality was not independent of ARDS subphenotype. Since serum creatinine was not associated with mortality in models (with or without subphenotype) (Table 2 and Table 4) and was independently associated with subphenotype (Table 3), it was apparent that cystatin C concentrations did not simply recapitulate findings from traditional creatininebased measurements.
Table 4.
Adjusted associations between 28-day mortality and selected covariates, including latent class analysis subphenotype
| Variate | Odds ratio (95% CI), univariable | P | Odds ratio (95% CI), multivariable | P |
|---|---|---|---|---|
| Age | 1.03 (1.02–1.04) | < 0.001 | 1.03 (1.01–1.04) | < 0.001 |
| Pao2:Fio2 ratio | 1.00 (0.99–1.00) | 0.368 | 1.00 (0.99–1.00) | 0.563 |
| Vasopressor used⁎ | 2.00 (1.26–3.25) | 0.004 | 1.59 (0.95–2.71) | 0.085 |
| Ln cystatin C† | 1.39 (1.12–1.72) | 0.002 | 1.16 (0.85–1.57) | 0.346 |
| Hyperinflammatory subphenotype‡ | 2.61 (1.69–4.02) | < 0.001 | 2.06 (1.16–3.64) | 0.013 |
| Serum creatinine | 1.14 (0.93–1.40) | 0.198 | 0.80 (0.58–1.07) | 0.146 |
| Treated with simvastatin§ | 0.68 (0.44–1.04) | 0.078 | 0.70 (0.44–1.11) | 0.134 |
Fio2 = fraction of inspired oxygen; Pao2 = partial pressure of arterial oxygen.
Reference category: vasopressor not used.
Natural logarithm-transformed baseline plasma cystatin C concentration.
Reference category: hypoinflammatory subphenotype.
Reference category: placebo.
It has previously been reported in the HARP-2 cohort that hyperinflammatory ARDS is predictive of a treatment response to simvastatin (ie, that patients with hyperinflammatory ARDS receive a mortality benefit from simvastatin treatment).10 As expected, we reproduced this finding in our 466-patient cohort from HARP-2 (OR, 2.05 [95% CI, 1.16–3.62]; P = 0.014) (Table 5). However, cystatin C concentration did not have statistically significant predictive power regarding response to simvastatin when included in a model with LCA subphenotype (OR, 1.10 [95% CI, 0.77–1.56]; P = 0.614) (Table 5) or without LCA subphenotype (OR, 1.20 [95% CI, 0.86–1.70]; P = 0.284) (Online Appendix, table 2). This was also the case for serum creatinine concentration.
Table 5.
Adjusted associations with mortality benefit following simvastatin treatment
| Variate | Odds ratio (95% CI), univariable | P | Odds ratio (95% CI), multivariable | P |
|---|---|---|---|---|
| Age | 1.03 (1.02-1.04) | < 0.001 | 1.03 (1.01–1.04) | 0.001 |
| Pao2:Fio2 ratio | 1.00 (0.99-1.00) | 0.368 | 1.00 (0.99–1.00) | 0.592 |
| Vasopressor used⁎ | 2.00 (1.26-3.25) | 0.004 | 1.58 (0.94–2.70) | 0.087 |
| Ln cystatin C† | 1.24 (0.94-1.63) | 0.128 | 1.10 (0.77–1.56) | 0.614 |
| Hyperinflammatory subphenotype‡ | 2.61 (1.69-4.02) | < 0.001 | 2.05 (1.16–3.62) | 0.014 |
| Serum creatinine | 1.14 (0.93-1.40) | 0.198 | 0.80 (0.58–1.07) | 0.151 |
| Treated with simvastatin§ | 0.69 (0.45-1.07) | 0.099 | 0.69 (0.43–1.09) | 0.116 |
| Simvastatin–cystatin C interaction | 1.28 (0.83-1.98) | 0.269 | 1.15 (0.72-1.86) | 0.558 |
Fio2 = fraction of inspired oxygen; Pao2 = partial pressure of arterial oxygen.
Reference category: vasopressor not used.
Natural logarithm-transformed baseline plasma cystatin C concentration.
Reference category: hypoinflammatory subphenotype.
Reference category: placebo. Interaction between simvastatin treatment and natural logarithm-transformed baseline plasma cystatin C concentration.
Discussion
In this retrospective analysis, we found that plasma cystatin C concentrations at baseline in the HARP-2 cohort are higher in patients with ARDS who do not survive past 28 days. This result, along with the finding that cystatin C concentrations are higher in patients who have hyperinflammatory ARDS, is largely in agreement with previously published data from another large cohort of patients with ARDS.6 However, in our cohort we found that, when included in multivariate models, cystatin C does not appear to add significantly to the prognostic power of LCA subphenotype. We therefore suggest that cystatin C concentration is not an independent predictor of mortality in the HARP-2 cohort. Furthermore, cystatin C concentration, unlike LCA subphenotype, cannot be used to predict response to treatment with simvastatin. Therefore, it appears that hyperinflammatory subphenotype is a good prognostic marker for mortality, while any prognostic power of cystatin C may be attributable to an association with subphenotype.
This study had some limitations. The modest number of patients with high serum creatinine concentrations in our analysis limited our ability to fully address the prognostic value of cystatin C. Although we report that cystatin C concentrations are higher in non-survivors with serum creatinine concentrations < 170 μmol/L, we are unable to draw a conclusion on the relationship between cystatin C concentration and mortality in patients with high serum creatinine concentrations. Therefore, our findings should be corroborated in larger cohorts of patients with ARDS. The reasons for an association between cystatin C concentration and mortality (and indeed LCA subphenotype) remain elusive and cannot be established from this work because of its retrospective nature. It has been proposed that cystatin C is simply a more sensitive marker of renal dysfunction than serum creatinine.6 Thus, elevated cystatin C concentrations may be a by-product of more advanced renal and multiorgan failure. In our study, we identified that cystatin C concentration but not serum creatinine concentration was associated with mortality, while both were independently associated with inflammatory subphenotype. These findings suggest that cystatin C is more strongly prognostic for mortality than creatinine and may therefore be a more sensitive marker for severity of disease and multiorgan failure. Risk of multiorgan failure is known to increase with increasing ARDS severity and is predictive of poorer outcomes,13, 14 so an early predictor of multiorgan dysfunction is very likely to have some prognostic value. Considering the robust association between cystatin C concentrations and hyperinflammatory subphenotype in our analysis, it is noteworthy that creatinine and bilirubin (traditional markers of renal and hepatic failure, respectively) have been shown to contribute to the hyperinflammatory subphenotype in several previously examined ARDS cohorts.8, 9, 10 The contributions made by markers of physiological dysfunction to the definition of LCA subphenotype have generally been less profound than those of inflammatory markers such as interleukin-6, interleukin-8 and soluble tumour necrosis factor receptor 1.8, 9, 10 Since cystatin C levels are thought to be affected by inflammation, it is possible that upregulation of cystatin C in ARDS is not just a consequence of early multiorgan impairment.7, 15, 16, 17 However, our study did not enable us to investigate this possibility.
What is also unclear is whether cystatin C has a role in the pathogenesis of ARDS. For example, higher levels of cystatin C may combat pathogenic cysteine proteases, which are a feature of many inflammatory conditions.18, 19, 20, 21, 22, 23 There is also an increasing awareness of the role of antiproteases beyond their protease inhibitory functions, and cystatin C is known to be involved in functions as diverse as antagonism of the transforming growth factor-β signal pathway, amyloidogenesis, and nuclear factor κ-B activity induced by interferon-γ.24, 25, 26 Therefore, further research into the control and biological function of cystatin C in ARDS is warranted to inform our understanding of its role as a biomarker and/or effector.
Conclusions
In this retrospective analysis of the HARP-2 cohort of patients with ARDS, we found that plasma cystatin C concentrations may be dependent on LCA subphenotype, potentially limiting the usefulness of cystatin C as a singular biomarker for ARDS mortality. More fundamental research is required to elucidate the reasons for elevated cystatin C concentrations in patients who do not survive and those with the hyperinflammatory subphenotype of ARDS.
Competing interests
Daniel McAuley has received personal fees from Bayer, GSK, Boehringer Ingelheim, Novartis, Eli Lilly, Vir Biotechnology, Faron Pharmaceuticals and Sobi, and is a named inventor on patent number US8962032B2. None of these relate directly to the study described in this article. All other authors declare that they have no competing interests.
Acknowledgements
Funding for this study was provided by the UK Medical Research Council (MR/P022847/1), the US National Institutes of Health (R35HL140026) and the Northern Ireland Department for the Economy (studentship for Michael McKelvey).
Footnotes
The whole patient cohort was divided into survivors and non-survivors at 28 days after enrolment (panel A). Patients were then further stratified by serum creatinine concentration (panel B) and latent class analysis subphenotype (panel C). Median values are represented by the horizontal bar in each box, interquartile ranges by the upper and lower limits of each box, ranges by whiskers, and outliers by individual points beyond the whiskers. t Natural logarithm-transformed baseline plasma cystatin C concentration.
Supplementary Information
References
- 1.Newman D.J., Cystatin C. Ann Clin Biochem. 2002;39:89–104. doi: 10.1258/0004563021901847. [DOI] [PubMed] [Google Scholar]
- 2.Zi M., Xu Y. Involvement of cystatin C in immunity and apoptosis. Immunol Lett. 2018;196:80–90. doi: 10.1016/j.imlet.2018.01.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Chew J.S.C., Saleem M., Florkowski C.M., George P.M. Cystatin C — a paradigm of evidence based laboratory medicine. Clin Biochem Rev. 2008;29:47–62. [PMC free article] [PubMed] [Google Scholar]
- 4.Benoit S.W., Ciccia E.A. Devarajan P Cystatin C as a biomarker of chronic kidney disease: latest developments. Expert Rev Mol Diagn. 2020;20:1019–1026. doi: 10.1080/14737159.2020.1768849. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bell M., Granath F., Mårtensson J., et al. Cystatin C is correlated with mortality in patients with and without acute kidney injury. Nephrol Dial Transplant. 2009;24:3096–3102. doi: 10.1093/ndt/gfp196. [DOI] [PubMed] [Google Scholar]
- 6.Hendrickson C.M., Kwong Y.D., Belzer A.G., et al. Higher plasma cystatin C is associated with mortality after acute respiratory distress syndrome: findings from a Fluid and Catheter Treatment Trial (FACTT) substudy. Crit Care. 2020;24:416. doi: 10.1186/s13054-020-03111-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chen D., Sun W., Li J., et al. Serum cystatin C and coronavirus disease 2019: a potential inflammatory biomarker in predicting critical illness and mortality for adult patients. Mediators Inflamm. 2020;2020:1–10. doi: 10.1155/2020/3764515. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Calfee C.S., Delucchi K., Parsons P.E., et al. Subphenotypes in acute respiratory distress syndrome: latent class analysis of data from two randomised controlled trials. Lancet Respir Med. 2014;2:611–620. doi: 10.1016/S2213-2600(14)70097-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sinha P., Delucchi K.L., Thompson B.T., et al. Latent class analysis of ARDS subphenotypes: a secondary analysis of the statins for acutely injured lungs from sepsis (SAILS) study. Intensive Care Med. 2018;44:1859–1869. doi: 10.1007/s00134-018-5378-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Calfee C.S., Delucchi K.L., Sinha P., et al. Acute respiratory distress syndrome subphenotypes and differential response to simvastatin: secondary analysis of a randomised controlled trial. Lancet Respir Med. 2018;6:691–698. doi: 10.1016/S2213-2600(18)30177-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.McAuley D.F., Laffey J.G., O’Kane C.M., et al. Simvastatin in the acute respiratory distress syndrome. N Engl J Med. 2014;371:1695–1703. doi: 10.1056/NEJMoa1403285. [DOI] [PubMed] [Google Scholar]
- 12.McAuley D.F., Laffey J.G., O’Kane C.M., et al. Hydroxymethylglutaryl-CoA reductase inhibition with simvastatin in acute lung injury to reduce pulmonary dysfunction (HARP-2) trial: study protocol for a randomized controlled trial. Trials. 2012;13:170. doi: 10.1186/1745-6215-13-170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Doyle R.L., Szaflarski N., Modin G.W., et al. Identification of patients with acute lung injury: predictors of mortality. Am J Respir Crit Care Med. 1995;152:1818–1824. doi: 10.1164/ajrccm.152.6.8520742. [DOI] [PubMed] [Google Scholar]
- 14.Kallet R.H., Lipnick M.S., Zhuo H., et al. Characteristics of nonpulmonary organ dysfunction at onset of ARDS based on the Berlin definition. Respir Care. 2019;64:493–501. doi: 10.4187/respcare.06165. [DOI] [PubMed] [Google Scholar]
- 15.Okura T., Jotoku M., Irita J., et al. Association between cystatin C and inflammation in patients with essential hypertension. Clin Exp Nephrol. 2010;14:584–588. doi: 10.1007/s10157-010-0334-8. [DOI] [PubMed] [Google Scholar]
- 16.Xu Y., Schnorrer P., Proietto A., et al. IL-10 controls cystatin C synthesis and blood concentration in response to inflammation through regulation of IFN regulatory factor 8 expression. J Immunol. 2011;186:3666–3673. doi: 10.4049/jimmunol.1001934. [DOI] [PubMed] [Google Scholar]
- 17.Duan Y., Xu J., Hu W., et al. The relevance of serum cystatin C level of different classification of atrial fibrillation. Int J Clin Med. 2019;10:404–412. [Google Scholar]
- 18.Vidak E., Javoršek U., Vizovišek M., Turk B. Cysteine cathepsins and their extracellular roles: shaping the microenvironment. Cells. 2019;8:264. doi: 10.3390/cells8030264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wartenberg M., Andrault P.-M., Saidi A., et al. Oxidation of cathepsin S by major chemicals of cigarette smoke. Free Radic Biol Med. 2020;150:53–65. doi: 10.1016/j.freeradbiomed.2020.02.013. [DOI] [PubMed] [Google Scholar]
- 20.Small D.M., Brown R.R., Doherty D.F., et al. Targeting of cathepsin S reduces cystic fibrosis-like lung disease. Eur Respir J. 2019;53:1801523. doi: 10.1183/13993003.01523-2018. [DOI] [PubMed] [Google Scholar]
- 21.Jobs E., Risérus U., Ingelsson E., et al. Serum cathepsin S is associated with serum C-reactive protein and interleukin-6 independently of obesity in elderly men. J Clin Endocrinol Metab. 2010;95:4460–4464. doi: 10.1210/jc.2010-0328. [DOI] [PubMed] [Google Scholar]
- 22.Martin S.L., Moffitt K.L., McDowell A., et al. Association of airway cathepsin B and S with inflammation in cystic fibrosis. Pediatr Pulmonol. 2010;45:860–868. doi: 10.1002/ppul.21274. [DOI] [PubMed] [Google Scholar]
- 23.McKelvey M.C., Abladey A.A., Small D.M., et al. Cathepsin S contributes to lung inflammation in acute respiratory distress syndrome. Am J Respir Crit Care Med. 2022;205:769–782. doi: 10.1164/rccm.202107-1631OC. [DOI] [PubMed] [Google Scholar]
- 24.Sokol J.P., Schiemann W.P. Cystatin C antagonizes transforming growth factor beta signaling in normal and cancer cells. Mol Cancer Res. 2004;2:183–195. [PubMed] [Google Scholar]
- 25.Perlenfein T.J., Mehlhoff J.D., Murphy R.M. Insights into the mechanism of cystatin C oligomer and amyloid formation and its interaction with -amyloid. J Biol Chem. 2017;292:11485–11498. doi: 10.1074/jbc.M117.786558. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Frendéus K.H., Wallin H., Janciauskiene S., Abrahamson M. Macrophage responses to interferon- are dependent on cystatin C levels. Int J Biochem Cell Biol. 2009;41:2262–2269. doi: 10.1016/j.biocel.2009.05.005. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials