Table 7 Studies measuring lung-specific proteins in patients with or at risk for ARDS
From: Serum biomarkers in Acute Respiratory Distress Syndrome an ailing prognosticator
Investigator | Patients Controls | Biomarker / Summary | ROC curve analysis Cut-off values | Specificity – Sensitivity Diagnostic accuracy | Limitations |
|---|---|---|---|---|---|
Doyle et al. 81 | 15 ARDSa 10 at risk 10 controls | SP-A is an acute indicator of lung function and alveolocapillary membrane injury | No | Not estimated | Small number of patients No ROC curve analysis / cut-off values No definitive relation with disease severity |
Doyle et al. 82 | 22 ARDSa 10 at risk 33 controls | Superiority of SP-B compared to SP-A plasma levels as a marker of lung function and alveolocapillary membrane injury | No | Not estimated | Only 3 case-control studies Inadequate sample size Lack of adjustment with disease behaviour No ROC curve analysis / cut-off values |
Greene et al. 83 | 41 ARDSa 22 at risk 35 controls | SP-A, SP-B, SP-D Serum changes found to be neither sensitive nor specific in predicting the onset of ARDS and discriminating survivors from non-survivors. | Yes Not reported | Poor predictive value Low specificity/sensitivity | Limited number of patients Serial measurements for a short period of time/ Lack of serial measurement for the most severe forms Heterogeneity of studied population Poor predictive value for serum levels |
Cheng et al. 84 | 36 ARDSa 2 ALI | SP-A levels were associated with severity of clinical lung injury and with disease outcome | No | Not estimated | Small sample size Causal diversity of studied population No serial measurement |
Greene et al. 85 | 51 at risk 26 ARDSa 16 controls | SP-A levels are predictive for at risk patients who developed ARDS from sepsis and aspiration but not trauma | No | Not estimated | Small sample size No ROC curve analysis / cut-off levels |
Bersten et al. 86 | 54 at risk 9 controls | SP-B but not SP-A cut-off plasma levels predict ARDS development, particularly in at-risk patients suffering a direct lung injury | Yes SP-B: 4.994 ng/ml | 78-85-85-78% | Small number of patients Limited follow-up serum data Most of patients had already lung injury Exclusion of milder at risk patients |
Eisner et al. 87 | 565 ALI/ ARDSa | SP-A, SP-D Attenuation of SP-D plasma levels by lower volume ventilation strategies | No | Not estimated | Only 2 serial measurements Heterogeneity of studied population Potential selection bias No ROC curve analysis / cut-off levels |
Ishizaka et al. 95 | 35 at risk 27 ARDSa 21 controls | Association of optimal cut-off values of KL-6 serum levels with patients' mortality | Yes KL-6: 253 U/ml | 100-87% | Inadequate sample volume Heterogeneity of studied population |
Sato et al. 96 | 28 ARDSa 10 controls | Association of KL-6 serum levels with variables of lung injury severity and with mortality rates No correlation with ventilation strategies | No | Not estimated | Small sample size Heterogeneity of studied group No serial measurement No ROC curve analysis / cut-off levels Diversity of ventilatory treatment |