Abstract
Objective
This study aims to review relevant research and assess the diagnostic value of serum cystatin C (CysC) for post-cardiac surgery acute kidney injury (PCSAKI).
Method
We searched databases (PubMed, Embase, Cochrane, WanFang, CNKI, VIP) for literature published up to January 10, 2024. Quality was assessed using Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2). Extracted data from eligible studies and summarized sensitivity, specificity, and area under the curve (AUC) values.
Results
A total of 24 studies involving 3,427 patients were included. The estimated diagnostic sensitivity of CysC for PCSAKI was 0.67 (95% CI, 0.57–0.76), with a specificity of 0.87 (95% CI, 0.81–0.91). The positive likelihood ratio (+LR) was 5.17 (95% CI, 3.45–7.73), and the negative likelihood ratio (-LR) was 0.38 (95% CI, 0.28–0.51). The diagnostic odds ratio (DOR) was 14 (95% CI, 7–26), the diagnostic score (DS) was 2.62 (95% CI, 1.99–3.24), and AUC was 0.86 (95% CI, 0.83–0.89). The sub-analysis results indicate that gender distribution, serum storage temperature, CysC detection method, and detection time all have a significant impact on sensitivity and specificity.
Conclusion
CysC has high specificity and good sensitivity in diagnosing PCSAKI during the perioperative period, with better detection results 24 hours before surgery, making it suitable for early detection. However, whether and how CysC is commonly used in clinical diagnosis still requires further research and clinical trials.
Introduction
Acute kidney injury (AKI) is a sudden loss of kidney function, characterized by elevated serum creatinine and blood urea nitrogen levels. It commonly occurs after cardiac surgery with cardiopulmonary bypass, with an incidence of 30%-40%. The duration and severity of AKI are closely linked to poor patient outcomes, increasing morbidity and mortality. Even with recovery of kidney function, some patients may develop subclinical or subacute kidney disease (AKD), leading to progressive renal disease [1–3].
Despite the advances in AKI research, current clinical diagnostic standards remain controversial, primarily relying on elevated serum creatinine or reduced urine output [4–6]. However, these standards are influenced by non-renal factors such as muscle mass and surgical environment, often leading to delayed and inaccurate diagnoses [7]. Therefore, there is an urgent need for more effective diagnostic markers for acute kidney injury, especially post-cardiac surgery.
Cystatin C (CysC) has been found to be a crucial predictive biomarker for detecting acute kidney injury following cardiac surgery [8]. It is a 13-kDa endogenous cysteine protease inhibitor produced by nucleated cells at a constant rate, freely filtered in the glomerulus, and metabolized without increases due to urinary tract infections or chronic non-renal diseases like malignancies [9]. Thus, its serum concentration provides a highly sensitive estimate of glomerular filtration rate [10].
So far, despite several studies using serum CysC to predict acute kidney injury after cardiac surgery, its effectiveness remains controversial due to differences in detection time, methods, sample sizes, etc. [11]. Therefore, this study aims to systematically evaluate the diagnostic value of serum CysC for PCSAKI through meta-analysis. We opt for a bivariate model as our primary analytical approach to enhance accuracy and minimize heterogeneity effects [12].
Methods
This meta-analysis was carried out according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [13].
Data sources and search strategy
Conducted a comprehensive search for literature published on PubMed, Embase, Cochrane Library, Wanfang, CNKI, and VIP databases up to January 10, 2024. We employed a search strategy to identify all trials using the following keywords: serum cystatin C or CysC, post-cardiac surgery, acute kidney injury, or renal failure. (see S1 Table for details). Additionally, we scanned the reference lists of all included studies and relevant reviews. The search was independently conducted by two researchers.
Study selection
We included all retrieved articles without restrictions on sample size or language. Initially, studies meeting the following criteria were identified:(1) Included articles with prospective cohort, case-control, or cross-sectional designs exploring serum CysC performance in PCSAKI detection. (2) Encompassed studies allowing the calculation of estimated sensitivity and specificity of serum CysC in diagnosing PCSAKI.
Two reviewers checked titles and abstracts of all citations using EndNote, followed by a full-text retrieval and reevaluation. The reference lists of reviewed articles were examined to ensure no omission of relevant studies. Any discrepancies were resolved by a third researcher.
Data extraction
A reviewer used a standardized form to extract information from each eligible study. The following information was extracted: (1) Study details: first author, publication year, country of origin, study design, sample size, and population characteristics; (2) Characteristics of study subjects: age, gender, baseline serum creatinine; (3) AKI information: AKI criteria and the number of AKI patients; (4) Serum CysC: measurement timing, measurement method, and optimal cutoff value; (5) Outcome information: sensitivity and specificity, true positive (TP), true negative (TN), false positive (FP), false negative (FN), and AUC value. If a study proposed more than one cutoff threshold, we used the threshold with the highest product of sensitivity and specificity.
Quality assessment of evidence
The methodological quality of these studies was independently assessed by two authors. We utilized the QUADAS-2 tool [14] to evaluate the quality of each trial. Risk of bias comprises four domains: patient selection, index test, reference test, and flow and timing. Applicability concerns encompass three domains: patient selection, index test, and reference test. Any discrepancies were resolved by a third researcher.
Data synthesis and analysis
We used a bivariate model for diagnostic test studies, offering stability with minimal additional information. This approach is especially valuable in situations with limited data, where the prior for the covariance matrix of the bivariate structure plays a crucial role, enhancing precision sensitivity and specificity, along with all hyperparameters and covariates, were accurately derived directly from the bivariate model, eliminating the need for Markov Chain Monte Carlo (MCMC) sampling [12]. Moreover, univariate estimates of sensitivity and specificity, with 95% confidence intervals (CIs), as well as the aggregated summary receiver operating characteristic (SROC) curve, can all be directly utilized for interpretation.
Calculate the 95% CI for the AUC of the SROC. Summarized +LR and -LR are derived from the summary estimates of sensitivity and specificity, respectively. We also assessed funnel plot asymmetry to effectively evaluate the degree and impact of publication and selective reporting bias in diagnostic accuracy studies. Subgroup analysis was performed considering study design, age range, gender ratio, sample storage temperature, diagnostic and testing criteria, and testing time.
Results
Search results
The initial search identified 4,277 records. After removing 523 studies due to duplicates, and further excluding 907 studies using automated tools, we screened the titles and abstracts of the remaining 2,811 studies. 44 studies underwent full-text review, with 24 ultimately included in the meta-analysis [15–38] (Fig 1).
Fig 1. PRISMA flowchart of the literature search.
Study and patient characteristics
This study included 24 studies with a total of 3427 patients. Technical features of individual studies are in Table 1. All these studies, published between 2008 and 2023, come from diverse countries and feature varied research designs, sample sizes (ranging from 40 to 628 individuals), and population settings. It’s worth noting that these articles use different definitions for AKI. The meta-analysis includes 22 prospective cohort studies and 2 retrospective case-control studies. Twenty studies used immunofluorescence to measure serum cystatin C levels, and the other four used commercial ELISA.
Table 1. Characteristics of studies included in the meta-analysis.
| Study | Country | Research type | Patients | Sample | CysC Assay | Baseline SCr (mg/dL) | AKI criteria | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| type | age(y) | Men(%) | AKI Diabetes(%) | size(AKI) | source | Storage(°C) | ||||||
| Koyner J 2008 | United States | Prospective study | Post-cardiac surgery | 63±13 | 70.8 | 21% | 72(34) | Serum | -80 | ELISA kit | 1.25±NA | RIFLE |
| Haase M 2009 | Australia | Prospective study | Post-cardiac surgery | 69.55±11.55 | 61 | 59% | 100(46) | Serum | -80 | Immunonephelometric | NA | AKIN |
| Haase-Fielitz 2009 | Australia | Prospective study | Post-cardiac surgery | 70.1±12.1 | 61.54 | 43% | 73(1)) | Serum | -80 | Immunonephelometric | NA | RIFLE |
| Che M 2010 | China | Retrospective study | Post-cardiac surgery | 62.9±13.7 | 68.96 | NA | 29(14) | Serum | -80 | Immunonephelometric | 0.827±0.135 | RIFLE-R |
| Krawczeski C 2010 | United States | Prospective study | Post-cardiac surgery | 4.2±5.6 | 53.47 | NA | 374(119) | Serum | -70 | Immunonephelometric | 0.435±0.235 | AKIN |
| Wald R 2010 | Canada | Prospective study | Post-cardiac surgery | 67±11.8 | 68.67 | 36% | 150(47) | Serum | -80 | Immunonephelometric | 1±0.1 | AKIN |
| Ristikankare A 2010 | Finland | Prospective study | Post-cardiac surgery | 77±3 | 54.54 | 34% | 110(62) | Serum | NA | Immunonephelometric | NA | RIFLE |
| Seitz S 2013 | Germany | Prospective study | Post-cardiac surgery | 1d to 44.3y | 54.68 | NA | 139(76) | Serum | -80 | Immunonephelometric | 0.38±0.21 | RIFLE |
| Yu CJ 2013 | China | Prospective study | Post-cardiac surgery | 61±12 | 40.5 | NA | 79(34) | Serum | -80 | Immunonephelometric | 0.773±0.264 | AKIN |
| Liebetrau C 2013 | Germany | Prospective study | Post-cardiac surgery | 69.5±12.2 | 68.79 | 33% | 141(47) | Serum | -80 | Immunonephelometric | 0.919±0.334 | KDIGO |
| Peco-Antić A 2013 | Serbia | Prospective study | Post-cardiac surgery | 1.6±3.19 | 58.03 | NA | 112(18) | Serum | -80 | ELISA kit | 0.6±0.124 | RIFLE |
| Zheng JY 2013 | China | Prospective study | Post-cardiac surgery | 12.55±13.95 | 69.77 | NA | 43(21) | Serum | -80 | Immunonephelometric | 0.362±0.169 | AKIN |
| Magro MC 2013 | Brazil | Prospective study | Post-cardiac surgery | 50±11 | 61.2 | NA | 121(16) | Serum | NA | Immunonephelometric | NA | RIFLE |
| Prowle JH 2015 | Australia | Prospective study | Post-cardiac surgery | 70±11.11 | 69 | 8% | 93(25) | Serum | -70 | Immunonephelometric | 1.086±0.31 | RIFLE |
| Hu XH 2015 | China | Prospective study | Post-cardiac surgery | 4.69±1.23 | 57 | NA | 100(26) | Serum | -40 | Immunonephelometric | NA | AKIN |
| Yong ZZ 2017 | China | Retrospective study | Post-cardiac surgery | 65.5±5.5 | 78.8 | 38% | 495(91) | Serum | NA | Immunonephelometric | 0.91±0.13 | AKIN |
| Kararmaz A 2019 | Turkey | Prospective study | Post-cardiac surgery | 56.9±20.2 | 21.4 | NA | 42(22) | Serum | NA | Immunonephelometric | 0.855±0.335 | KDIGO |
| Wang XD 2020 | China | Prospective study | Post-cardiac surgery | 63±11.85 | 64.17 | 24% | 628(178) | Serum | -80 | Immunonephelometric | 0.807±0.235 | KDIGO |
| Zheng XF 2021 | China | Prospective study | Post-cardiac surgery | 57±16.5 | 60 | 23% | 145(90) | Serum | -80 | Immunonephelometric | 1.25±0.4 | AKIN |
| Szymanowicz W 2021 | Poland | Prospective study | Post-cardiac surgery | 69±13 | 50 | 39% | 114(18) | Serum | -70 | Immunonephelometric | 1±0.29 | KDIGO |
| Lakhal K 2021 | France | Prospective study | Post-cardiac surgery | 78.5±8.15 | 49.2 | 19% | 65(27) | Serum | -80 | Immunonephelometric | NA | KDIGO |
| Kalisnik JM 2022 | Germany | Prospective study | Post-cardiac surgery | 73.1±11.1 | 36.97 | 22% | 119(51) | Serum | -20 | ELISA kit | 0.785±0.335 | KDIGO |
| Zakaria M 2022 | Egypt | Prospective study | Post-cardiac surgery | 3.65±NA | 52.5 | NA | 40(12) | Serum | NA | Immunonephelometric | 0.8±0.12 | AKIN |
| Abadeer M 2023 | United States | Prospective study | Post-cardiac surgery | 3.7±4.3m | 67.44 | NA | 43(18) | Serum | -80 | ELISA kit | NA | KDIGO |
Note: NA, Not Available
Quality assessment
The methodological quality of the studies according to the QUADAS-2 is summarized in Fig 2 and Table 2. The literature included in this study exhibits an overall low risk of bias and applicability concerns.
Fig 2. Overall level of risk of bias and applicability concerns.
Table 2. Quality assessment of the included studies.
| Studys | Risk of bias | Applicability concerns | |||||
|---|---|---|---|---|---|---|---|
| patient selection | Index test | Reference standard | Flow and timing | patient selection | Index test | Reference standard | |
| Koyner J 2008 | High | Low | Unclear | Low | Low | Unclear | Unclear |
| Haase M 2009 | Unclear | Unclear | Low | Low | Low | Low | Low |
| Haase-Fielitz 2009 | Unclear | High | Unclear | Low | Low | Low | Unclear |
| Che M 2010 | High | Low | Unclear | Low | Low | Low | Unclear |
| Krawczeski C 2010 | Low | Low | Low | Low | Low | Low | Low |
| Wald R 2010 | Low | Unclear | Low | Low | Low | Low | Low |
| Ristikankare A 2010 | Low | Unclear | Unclear | Low | Low | Low | Unclear |
| Seitz S 2013 | Low | Low | Low | Unclear | Low | Low | Unclear |
| Yu CJ 2013 | Low | Unclear | Low | Low | Low | Low | Low |
| Liebetrau C 2013 | Low | Low | Low | Unclear | Low | Unclear | Low |
| Peco-Antić A 2013 | Low | Low | Unclear | Low | Low | Low | Unclear |
| Zheng JY 2013 | Low | High | Low | High | Low | Low | Low |
| Magro MC 2013 | Unclear | Low | Low | Low | Low | Low | Low |
| Prowle JH 2015 | Unclear | Low | Unclear | Low | Low | Unclear | Unclear |
| Hu XH 2015 | Low | Low | Low | Low | Low | Unclear | Low |
| Yong ZZ 2017 | Low | Low | Low | Unclear | Low | Low | Low |
| Kararmaz A 2019 | Low | Low | Low | Low | Low | Low | Low |
| Wang XD 2020 | Low | Unclear | Low | Low | Low | Unclear | Low |
| Zheng XF 2021 | Low | Low | Low | Unclear | Low | Unclear | Low |
| Szymanowicz W 2021 | Low | Low | Low | Low | Low | Unclear | Low |
| Lakhal K 2021 | Low | Unclear | Low | Low | Low | Unclear | Low |
| Kalisnik JM 2022 | Low | Unclear | Low | Low | Low | Low | Low |
| Zakaria M 2022 | Low | Low | Low | Low | Low | Low | Low |
| Abadeer M 2023 | Low | Low | Low | Unclear | Low | Low | Low |
Data synthesis
Data from 24 eligible studies, detailed in Table 3, include TP, FN, FP, and TN. The information covers Cysc measurement methods, timing, optimal cutoff values, as well as sensitivity (95% CI), specificity (95% CI), and AUC-ROC (95% CI). Serum CysC showed a sensitivity of 0.67 (95% CI, 0.57–0.76) and specificity of 0.87 (95% CI, 0.81–0.91) in diagnosing PCSAKI (Fig 3). The +LR was 5.17 (95% CI, 3.45–7.73), and the -LR was 0.38 (95% CI, 0.28–0.51) (Fig 4). The diagnostic score was 2.62 (95% CI, 1.99–3.24), with an odds ratio of 13.7 (95% CI, 7.35–25.58) (Fig 5).The cross-axis plot displays sensitivity, false positive rate values, and confidence intervals for each included study. The SROC curve indicates that serum CysC1 exhibits high efficiency in the diagnosis of PCSAKI, with an AUC of 0.86 (95% CI, 0.83–0.89) (Fig 6). The Fagan plot demonstrates a significant improvement in post-test probability compared to pre-test probability when using CysC as a biomarker (Fig 7). Furthermore, a good consistency of results is indicated by the proximity of the peak to the coordinates (1, 1) in the posterior density distribution plot and a funnel plot p-value above 0.05 indicates no bias risk (Fig 8).
Table 3. Serum CysC performance in PCSAKI diagnosis.
| Studys | Patients | CysC | Sensitivity(%) | Specificity(%) | AUROC(95%CI) | ||||
|---|---|---|---|---|---|---|---|---|---|
| TP | FP | FN | TN | Time of Measurement | Cutoff Value | ||||
| Koyner J 2008 | 22 | 12 | 15 | 23 | <24h after surgery | NA | 65 | 61 | 0.624(0.493–0.754) |
| Haase M 2009 | 34 | 18 | 12 | 36 | <24h after surgery | 1.1mg/L | 74 | 67 | 0.76(0.61–0.91) |
| Haase-Fielitz 2009 | 12 | 3 | 6 | 52 | <24h before surgery | 1.1mg/L | 75 | 89 | 0.78(0.58–0.99) |
| Haase-Fielitz 2009(2) | 13 | 2 | 11 | 47 | <24h after surgery | 1.2mg/L | 86 | 80 | 0.84(0.75–0.93) |
| Che M 2010 | 10 | 1 | 4 | 14 | 10h after surgery | 1.31mg/L | 71 | 92 | 0.83(0.67–1) |
| Krawczeski C 2010 | 87 | 32 | 38 | 217 | 12h after surgery | 1.16mg/L | 73 | 85 | 0.81(0.74–0.88) |
| Wald R 2010 | 32 | 15 | 40 | 63 | 2h after surgery | NA | 67 | 61 | 0.68(0.58–0.78) |
| Ristikankare A 2010 | 42 | 20 | 19 | 29 | <24h after surgery | NA | 67 | 61 | 0.71(0.61–0.81) |
| Ristikankare A 2010(2) | 50 | 12 | 15 | 33 | <48h after surgery | NA | 79 | 68 | 0.77(0.68–0.86) |
| Seitz S 2013 | 61 | 34 | 2 | 42 | 2h after surgery | 0.995mg/L | 80 | 66 | 0.7401(0.606–0.875) |
| Yu CJ 2013 | 34 | 1 | 0 | 44 | 6h after surgery | NA | 100 | 98 | 0.95(NA) |
| Yu CJ 2013(2) | 34 | 0 | 0 | 45 | 48h after surgery | NA | 100 | 100 | 1(NA) |
| Liebetrau C 2013 | 42 | 5 | 30 | 64 | 4h after surgery | NA | 89 | 68 | 0.76(0.65–0.94) |
| Peco-AntićA 2013 | 15 | 3 | 40 | 54 | 2h after surgery | NA | 81.85 | 57.2 | 0.73(NA) |
| Zheng JY 2013 | 18 | 3 | 11 | 11 | 6h after surgery | 0.47mg/L | 85.7 | 50 | 0.69(NA) |
| Magro MC 2013 | 11 | 5 | 47 | 58 | <24h after surgery | NA | 66 | 55 | 0.719(0.62–0.817) |
| Magro MC 2013(2) | 24 | 10 | 35 | 52 | <24h after surgery | NA | 71 | 60 | 0.721(NA) |
| Prowle JHF 2015 | 19 | 6 | 25 | 43 | <24h after surgery | 1.24 | 76 | 63.2 | 0.69(0.56–0.82) |
| Prowle JH 2015(2) | 16 | 9 | 15 | 53 | <48h after surgery | 1.57 | 64 | 77.9 | 0.72(0/59-0.85) |
| Hu XH 2015 | 21 | 5 | 29 | 45 | <24h after surgery | NA | 80 | 60 | 0.7(NA) |
| Yong ZZ 2017 | 65 | 119 | 26 | 285 | 4-5d after surgery | 1.03mg/L | 71.43 | 70.54 | 0.77(0.71–0.82) |
| Kararmaz A 2019 | 21 | 2 | 1 | 18 | 2h after surgery | 0.85mg/L | 93 | 87 | 0.902(0.79–1) |
| Kararmaz A 2019(2) | 18 | 4 | 4 | 16 | 2h before surgery | 0.85mg/L | 78 | 80 | 0.796(NA) |
| Wang XD 2020 | 159 | 19 | 187 | 263 | 10h after surgery | 1.32mg/L | 88.8 | 58.3 | 0.75(NA) |
| Zheng XF 2021 | 78 | 16 | 12 | 39 | <24h after surgery | 0.975mg/L | 86.5 | 71.4 | 0.804(0.726–0.881) |
| Szymanowicz W 2021 | 16 | 2 | 5 | 91 | <24h after surgery | 1.23mg/L | 88 | 94 | 0.914(0.82–1) |
| Lakhal K 2021 | 14 | 13 | 14 | 24 | <24h after surgery | NA | 51 | 63 | 0.69(0.56–0.81) |
| Kalisnik JM 2022 | 32 | 19 | 33 | 35 | <24h after surgery | NA | 62 | 51 | 59(NA) |
| Zakaria M 2022 | 12 | 0 | 8 | 20 | <24h after surgery | 1.37mg/L | 100 | 68.8 | 0.781(0.58–0.97) |
| Zakaria M 2022(2) | 12 | 0 | 3 | 25 | <48h after surgery | 1.27mg/L | 1 | 87.5 | 1(1–1) |
| Abadeer M 2023 | 10 | 8 | 12 | 13 | <24h after surgery | NA | 56 | 52 | 0.54(NA) |
Note:NA,Not Available; TP: true positive; TN: true negative; FP: false positive; FN: false negative.
Fig 3.
Forest plots illustrating overall sensitivity (left) and specificity (right) of serum cystatin C in diagnosing PCSAKI.
Fig 4.
Forest plots illustrating overall +LR (left) and -LR (right) of serum cystatin C in diagnosing.
Fig 5.
Forest plots illustrating dagnostic score (left) and odds ratio (right) of serum cystatin C in diagnosing PCSAKI.
Fig 6.
Crosshair plots of the pooled sensitivity (left) and SROC curve (right) of serum cystatin C in diagnosing PCSAKI.
Fig 7. Fagan plot: A significant improvement in post-test probability compared to pre-test probability.
Fig 8.
Posterior Density Distribution Plot (Left) and Funnel Plot (Right): Assessing Potential Bias in the Diagnosis of PCSAKI Using Serum CysC.
This study found significant heterogeneity in combining sensitivity and specificity. Threshold analysis with a Spearman correlation coefficient of -0.077 and a p-value of 0.680 indicated that heterogeneity was not due to threshold effects. Subgroup analyses based on various criteria were then performed, and the results are presented in Table 4.
Table 4. Subgroup analysis using various criteria.
| studies | Sensitivity(95%Cl) | Specifcity(95%CI) | +LR (95%CI) | −LR (95%CI) | DOR(95%CI) | DS(95%CI) | AUC(95%CI) |
|---|---|---|---|---|---|---|---|
| All studies (31) | 0.67(0.57–0.76) | 0.87(0.81–0.91) | 5.17(3.45–7.73) | 0.38(0.28–0.51) | 14(7,26) | 2.62(1.99–3.24) | 0.86(0.83–0.89) |
| Study design | |||||||
| Prospective (29) | 0.67(0.56–0.76) | 0.87(0.82–0.91) | 5.3(3.4–8.1) | 0.38(0.27–0.52) | 14(7,27) | 2.64(1.97–3.3) | 0.86(0.83–0.89) |
| Retrospective study(2) | 0.71(0.62–0.80) | 0.73(0.68–0.77) | 2.42(1.94–2.91) | 0.39(0.27–0.51) | 9.87(2.07–47.09) | 2.01(1.50–2.51) | NA |
| Age | |||||||
| Children(7) | 0.64(0.41–0.83) | 0.89(0.74–0.96) | 6(2.4–14.7) | 0.4(0.22–0.72) | 15(5–47) | 2.71(1.56–3.88) | 0.86(0.83–0.89) |
| adult(24) | 0.68(0.57–0.78) | 0.86(0.80–0.91) | 5.0(3.2–8.0) | 0.37(0.26–0.52) | 14(6–29) | 2.62(1.87–3.37) | 0.86(0.83–0.89) |
| Sex ratio | |||||||
| Men>50%(24) | 0.61(0.52–0.69) | 0.85(0.79–0.89) | 4.0(3.1–5.2) | 0.46(0.38–0.56) | 9(6–12) | 2.15(1.80–2.51) | 0.81(0.78–0.85) |
| Men≤50%(7) | 0.87(0.63–0.96) | 0.93(0.78–0.98) | 13.1(3.1–56.3) | 0.14(0.04–0.49) | 92(7–1252) | 4.52(1.92–7.13) | 0.96(0.94–0.97) |
| Sample Storage | |||||||
| -80°C(16) | 0.72(0.55–0.84) | 0.87(0.78–0.93) | 5.6(2.9–10.6) | 0.32(0.18–0.56) | 17(6–51) | 2.88(1.79–3.93) | 0.88(0.85–0.91) |
| <-80°C(6) | 0.56(0.45–0.67) | 0.88(0.78–0.94) | 4.6(2.2–9.6) | 0.50(0.37–0.68) | 9(3–25) | 2.21(1.21–3.21) | 0.76(0.72–0.80) |
| unclear(9) | 0.68(0.50–0.81) | 0.86(0.74–0.93) | 4.7(2.4–9.3) | 0.38(0.23–0.63) | 12(5–34) | 2.52(1.51–3.53) | 0.85(0.82–0.88) |
| Diagnostic criteria | |||||||
| RIFLE criteria(10) | 0.57(0.38–0.74) | 0.86(0.77–0.92) | 4.2(2.8–6.1) | 0.50(0.34–0.74) | 8(5–15) | 2.12(1.53–2.71) | 0.83(0.80–0.86) |
| AKIN criteria(9) | 0.72(0.55–0.84) | 0.84(0.74–0.91) | 4.6(2.5–8.4) | 0.34(0.20–0.57) | 14(5–39) | 2.62(1.59–3.65) | 0.86(0.83–0.89) |
| KDIGO Classification(12) | 0.71(0.55–0.83) | 0.90(0.77–0.96) | 6.9(2.6–18.3) | 0.33(0.19–0.56) | 21(5–90) | 3.05(1.61–4.50) | 0.87(0.84–0.90) |
| Assay method | |||||||
| ELISA kit(4) | 0.45(0.33–0.58) | 0.76(0.55–0.89) | 1.9(1.1–3.1) | 0.73(0.61–0.86) | 3(1–5) | 0.94(0.29–1.58) | 0.58(0.54–0.62) |
| Immunonephelometric assay(27) | 0.70(0.60–0.79) | 0.88(0.83–0.92) | 6.0(3.9–9.2) | 0.34(0.24–0.47) | 18(9–35) | 2.88(2.21–3.55) | 0.8(0.85–0.91) |
| Test time | |||||||
| <24h before surgery(2) | 0.77(0.63–0.90) | 0.93(0.87–1.00) | 2.31(1.39–3.23) | 0.32(0.16–0.48) | 2.02(1.48–2.56) | 5.90(0.40–11.41) | NA |
| <24h after surgery(24) | 0.63(0.51–0.73) | 0.87(0.81–0.91) | 4.8(3.2–7.3) | 0.43(0.32–0.58) | 11(6–21) | 2.42(1.79–3.04) | 0.84(0.81–0.87) |
| >24h after surgery(5) | 0.83(0.58–0.95) | 0.94(0.59–0.99) | 14.3(1.3–162.3) | 0.18(0.06–0.56) | 79(3–2336) | 4.38(0.99–7.76) | 0.94(0.92–0.96) |
| AKI Diabetes(%) | |||||||
| ≤25%(7) | 0.57(0.44–0.69) | 0.79(0.68–0.87) | 2.7(1.8–4.2) | 0.54(0.42–0.71) | 5(3–9) | 1.62(1.00–2.23) | 0.74(0.70–0.77) |
| >25%(9) | 0.65(0.57–0.72) | 0.86(0.74–0.93) | 4.6(2.4–8.6) | 0.40(0.33–0.50) | 11(5–24) | 2.43(1.69–3.17) | 0.76(0.72–0.81) |
Note:NA,Not Available; +LR,positive likelihood ratio; -LR,negative likelihood ratio; DOR,dagnostic odds ratio;DS,dagnostic score; AUC,area under the cu
The findings reveal that gender distribution, serum storage temperature, CysC detection method, and testing time have a substantial impact on sensitivity, specificity, and AUC area. In diagnostics, accuracy is higher in the female sample group, -80°C storage, and immunofluorescence testing. Testing more than 24 hours after surgery is superior to less than 24 hours before surgery, which, in turn, is better than less than 24 hours after surgery.
Discussion
This article reviews relevant studies and systematically evaluates the diagnostic value of serum CysC for PCSAKI. By including 24 studies with a total of 3427 patients, it was found that CysC has high specificity (0.87, 95% CI, 0.81–0.91) and moderate sensitivity (0.67, 95% CI, 0.57–0.76) in diagnosing PCSAKI during the perioperative period. The comprehensive analysis showed that the area under the SROC curve was 0.86 (95% CI, 0.83–0.89), indicating that serum CysC has a high predictive value for diagnosing PCSAKI during the perioperative period.
Early diagnosis of AKI is crucial for treatment and prognosis. Currently, PCSAKI is typically diagnosed by observing elevated serum creatinine levels or reduced urine output [39]. However, existing diagnostic standards fail to meet clinical needs, as the damage may already be irreversible by the time AKI is detected [40]. Therefore, there is an urgent need for more effective methods to diagnose AKI at an earlier stage.
To achieve the goal of early prediction of PCSAKI, we divided the detection times into three subgroups: 0–24 hours preoperatively, 0–24 hours postoperatively, and 24–72 hours postoperatively. We assessed the diagnostic performance at these different time points. The results showed that the detection at 24–72 hours postoperatively had the best performance: sensitivity of 0.83 (95% CI, 0.58–0.95) and specificity of 0.94 (95% CI, 0.59–0.99). The next best was 0–24 hours preoperatively: sensitivity of 0.77 (95% CI, 0.63–0.90) and specificity of 0.93 (95% CI, 0.87–1.00). The worst performance was 0–24 hours postoperatively: sensitivity of 0.63 (95% CI, 0.51–0.73) and specificity of 0.87 (95% CI, 0.81–0.91). Comprehensive analysis indicates that 0–24 hours postoperatively has the poorest diagnostic performance, which is not in line with the goal of early prediction. The diagnostic specificity of 0–24 hours preoperatively is similar to that of 24–72 hours postoperatively, with a narrower 95% confidence interval, indicating a more precise diagnosis, and the sensitivity is also similar. More importantly, the 0–24 hours preoperative detection allows for earlier identification of patients who may develop PCSAKI, enabling preparations for potential kidney injury. This allows medical staff to better manage potential AKI situations and save medical resources by focusing on patients who are predicted to develop PCSAKI.
To facilitate its clinical application, we analyzed the cut-off values used in the studies. The results showed that most of the studies had cut-off values concentrated around 1.2 mg/L, which could provide a reference for the establishment of future detection standards. However, the cut-off values in the included studies varied widely (ranging from 0.47 to 1.57 mg/L). Therefore, more research is needed to further standardize these values to achieve consistent diagnostic results in clinical practice.
In addition, we conducted a series of subgroup analyses to explore factors potentially influencing the CysC. We found that the sensitivity and specificity of CysC were influenced by gender, sample storage temperature, detection method, detection time, and the presence of diabetes. Firstly, our subgroup analysis revealed that CysC exhibited higher sensitivity and specificity in samples where males accounted for less than 50%. Literature reports also indicate gender differences in CysC, especially in children, which is consistent with our findings [41]. Secondly, sample storage temperature significantly affected CysC detection results. Samples stored at -70°C showed lower sensitivity and specificity compared to those stored at -80°C. This suggests that sample storage conditions should be strictly controlled in standardized detection processes. Additionally, we found that immunofluorescence detection methods have higher sensitivity and specificity relative to ELISA assay kits. Therefore, selecting appropriate detection methods is crucial for improving the diagnostic accuracy of CysC. It is noteworthy that we found a higher incidence of diabetes among PCSAKI patients, indicating that these patients are more prone to AKI, consistent with existing research findings [42, 43]. Additionally, in high-risk diabetic patients, there is a significant increase in the sensitivity and specificity of CysC detection, further supporting the close association between diabetes and PCSAKI. Therefore, we recommend considering the use of new antidiabetic medications during the perioperative period, especially those that do not harm kidney function and do not interfere with testing. This approach may have significant benefits in reducing the risk of kidney injury after cardiac surgery [44].
This study also assessed the consistency of included studies through density plot analysis, with results showing a peak close to (1, 1), indicating good consistency among the study outcomes. Funnel plot analysis revealed no publication bias, and overall quality assessment indicated high study quality. Considering potential threshold effects and the correlation between sensitivity and specificity, threshold effect analysis was conducted, which indicated no threshold effect. This suggests that the results of this study are reliable.
Strengths and limitations
This study integrated multiple research results, increased sample size, and improved the accuracy of CysC in diagnosing PCSAKI. It found that CysC can predict PCSAKI before surgery, enabling early intervention, and proposed diagnostic cutoff values for clinical use. The study also identified factors affecting CysC’s diagnosis of PCSAKI, offering guidance for precise diagnostics and ensuring result reliability through various quality control assessments.
However, studies included in this research have shown extreme variability in the diagnostic cutoff values of CysC, necessitating further clinical studies for validation and supplementation. Moreover, blind controls in some included studies were flawed, potentially affecting the objectivity of the results. Lastly, the conclusions are only applicable to patients after cardiac surgery; further research is needed to assess applicability to other patient populations.
Conclusion
Serum CysC demonstrates high specificity in the diagnosis of acute kidney injury following cardiac surgery, with high sensitivity when detected within 0–24 hours preoperatively, enabling early prediction. However, further research is needed to determine its practical application in clinical settings.
Registration number
This study has been registered with PROSPERO and the registration number is CRD42024521067.
Supporting information
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
Data Availability
All relevant data are within the paper and its Supporting Information files. Please see Tables 1 and 3, as well as the corresponding references (15-38).
Funding Statement
The author(s) received no specific funding for this work.
References
- 1.Kashani K., Cheungpasitporn W., and Ronco C., Biomarkers of acute kidney injury: the pathway from discovery to clinical adoption. Clin Chem Lab Med, 2017. 55(8): p. 1074–1089. doi: 10.1515/cclm-2016-0973 [DOI] [PubMed] [Google Scholar]
- 2.Farrar A., Acute Kidney Injury. Nurs Clin North Am, 2018. 53(4): p. 499–510. doi: 10.1016/j.cnur.2018.07.001 [DOI] [PubMed] [Google Scholar]
- 3.Mizuguchi K.A., et al., Predicting kidney disease progression in patients with acute kidney injury after cardiac surgery. J Thorac Cardiovasc Surg, 2018. 155(6): p. 2455–2463.e5. doi: 10.1016/j.jtcvs.2018.01.093 [DOI] [PubMed] [Google Scholar]
- 4.Bellomo R., et al., Acute renal failure—definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care, 2004. 8(4): p. R204–12. doi: 10.1186/cc2872 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mehta R.L., et al., Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care, 2007. 11(2): p. R31. doi: 10.1186/cc5713 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ricci Z. and Romagnoli S., Acute Kidney Injury: Diagnosis and Classification in Adults and Children. Contrib Nephrol, 2018. 193: p. 1–12. doi: 10.1159/000484956 [DOI] [PubMed] [Google Scholar]
- 7.Koyner J.L., et al., Biomarkers predict progression of acute kidney injury after cardiac surgery. J Am Soc Nephrol, 2012. 23(5): p. 905–14. doi: 10.1681/ASN.2011090907 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kiessling A.H., et al., Early postoperative serum cystatin C predicts severe acute kidney injury following cardiac surgery: a post-hoc analysis of a randomized controlled trial. J Cardiothorac Surg, 2014. 9: p. 10. doi: 10.1186/1749-8090-9-10 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang Y. and Bellomo R., Cardiac surgery-associated acute kidney injury: risk factors, pathophysiology and treatment. Nat Rev Nephrol, 2017. 13(11): p. 697–711. doi: 10.1038/nrneph.2017.119 [DOI] [PubMed] [Google Scholar]
- 10.Palevsky P.M., et al., KDOQI US commentary on the 2012 KDIGO clinical practice guideline for acute kidney injury. Am J Kidney Dis, 2013. 61(5): p. 649–72. doi: 10.1053/j.ajkd.2013.02.349 [DOI] [PubMed] [Google Scholar]
- 11.Yong Z., et al., Predictive value of serum cystatin C for acute kidney injury in adults: a meta-analysis of prospective cohort trials. Sci Rep, 2017. 7: p. 41012. doi: 10.1038/srep41012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Guo J., Riebler A., and Rue H., Bayesian bivariate meta-analysis of diagnostic test studies with interpretable priors. Stat Med, 2017. 36(19): p. 3039–3058. doi: 10.1002/sim.7313 [DOI] [PubMed] [Google Scholar]
- 13.Page M.J., et al., The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Bmj, 2021. 372: p. n71. doi: 10.1136/bmj.n71 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Whiting P.F., et al., QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med, 2011. 155(8): p. 529–36. doi: 10.7326/0003-4819-155-8-201110180-00009 [DOI] [PubMed] [Google Scholar]
- 15.Abadeer M., et al., Using Serum Cystatin C to Predict Acute Kidney Injury Following Infant Cardiac Surgery. Pediatr Cardiol, 2023. 44(4): p. 855–866. doi: 10.1007/s00246-022-03080-y [DOI] [PubMed] [Google Scholar]
- 16.Che M., et al., Clinical usefulness of novel biomarkers for the detection of acute kidney injury following elective cardiac surgery. Nephron Clin Pract, 2010. 115(1): p. c66–72. doi: 10.1159/000286352 [DOI] [PubMed] [Google Scholar]
- 17.Haase M., et al., Novel biomarkers early predict the severity of acute kidney injury after cardiac surgery in adults. Ann Thorac Surg, 2009. 88(1): p. 124–30. doi: 10.1016/j.athoracsur.2009.04.023 [DOI] [PubMed] [Google Scholar]
- 18.Haase-Fielitz A., et al., Novel and conventional serum biomarkers predicting acute kidney injury in adult cardiac surgery—a prospective cohort study. Crit Care Med, 2009. 37(2): p. 553–60. doi: 10.1097/CCM.0b013e318195846e [DOI] [PubMed] [Google Scholar]
- 19.Kalisnik J.M., et al., Enhanced Detection of Cardiac Surgery-Associated Acute Kidney Injury by a Composite Biomarker Panel in Patients with Normal Preoperative Kidney Function. J Cardiovasc Dev Dis, 2022. 9(7). doi: 10.3390/jcdd9070210 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kararmaz A., et al., Evaluation of acute kidney injury with oxidative stress biomarkers and Renal Resistive Index after cardiac surgery. Acta Chir Belg, 2021. 121(3): p. 189–197. doi: 10.1080/00015458.2019.1702371 [DOI] [PubMed] [Google Scholar]
- 21.Koyner J.L., et al., Urinary cystatin C as an early biomarker of acute kidney injury following adult cardiothoracic surgery. Kidney Int, 2008. 74(8): p. 1059–69. doi: 10.1038/ki.2008.341 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Krawczeski C.D., et al., Serum cystatin C is an early predictive biomarker of acute kidney injury after pediatric cardiopulmonary bypass. Clin J Am Soc Nephrol, 2010. 5(9): p. 1552–7. doi: 10.2215/CJN.02040310 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Lakhal K., et al., Early recognition of cardiac surgery-associated acute kidney injury: lack of added value of TIMP2 IGFBP7 over short-term changes in creatinine (an observational pilot study). BMC Anesthesiol, 2021. 21(1): p. 244. doi: 10.1186/s12871-021-01387-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Magro M.C. and Vattimo Mde F., Impact of cystatin C and RIFLE on renal function assessment after cardiac surgery. Biol Res Nurs, 2013. 15(4): p. 451–8. doi: 10.1177/1099800412446742 [DOI] [PubMed] [Google Scholar]
- 25.Prowle J.R., et al., Combination of biomarkers for diagnosis of acute kidney injury after cardiopulmonary bypass. Ren Fail, 2015. 37(3): p. 408–16. doi: 10.3109/0886022X.2014.1001303 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ristikankare A., et al., Serum cystatin C in elderly cardiac surgery patients. Ann Thorac Surg, 2010. 89(3): p. 689–94. doi: 10.1016/j.athoracsur.2009.11.018 [DOI] [PubMed] [Google Scholar]
- 27.Seitz S., et al., Cystatin C and neutrophil gelatinase-associated lipocalin: biomarkers for acute kidney injury after congenital heart surgery. Swiss Med Wkly, 2013. 143: p. w13744. doi: 10.4414/smw.2013.13744 [DOI] [PubMed] [Google Scholar]
- 28.Szymanowicz W., et al., Brain and Muscle Oxygen Saturation Combined with Kidney Injury Biomarkers Predict Cardiac Surgery Related Acute Kidney Injury. Diagnostics (Basel), 2021. 11(9). doi: 10.3390/diagnostics11091591 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wald R., et al., Plasma cystatin C and acute kidney injury after cardiopulmonary bypass. Clin J Am Soc Nephrol, 2010. 5(8): p. 1373–9. doi: 10.2215/CJN.06350909 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Wang X., et al., Early serum cystatin C-enhanced risk prediction for acute kidney injury post cardiac surgery: a prospective, observational, cohort study. Biomarkers, 2020. 25(1): p. 20–26. doi: 10.1080/1354750X.2019.1688865 [DOI] [PubMed] [Google Scholar]
- 31.Zheng J.Y., et al., Is serum cystatin C an early predictor for acute kidney injury following cardiopulmonary bypass surgery in infants and young children? Kaohsiung J Med Sci, 2013. 29(9): p. 494–9. doi: 10.1016/j.kjms.2013.01.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Liebetrau C., et al., Neutrophil gelatinase-associated lipocalin (NGAL) for the early detection of cardiac surgery associated acute kidney injury. Scand J Clin Lab Invest, 2013. 73(5): p. 392–9. doi: 10.3109/00365513.2013.787149 [DOI] [PubMed] [Google Scholar]
- 33.Peco-Antić A., et al., Biomarkers of acute kidney injury in pediatric cardiac surgery. Clin Biochem, 2013. 46(13–14): p. 1244–51. doi: 10.1016/j.clinbiochem.2013.07.008 [DOI] [PubMed] [Google Scholar]
- 34.Zakaria M., et al., Role of serum cystatin C in the prediction of acute kidney injury following pediatric cardiac surgeries: A single center experience. Medicine (Baltimore), 2022. 101(49): p. e31938. doi: 10.1097/MD.0000000000031938 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Hu XH, Niu Zl, Wang J, et al., Diagnostic value of kidney injury molecular 1, cystatin C and urinary creatinine in acute kidney injury induced by cardiopulmonary bypass of congenital heart disease. Hainan Medicine, 2015. 26(13): p. 1934–1937. [Google Scholar]
- 36.Yong ZZ, Application of serum cystatin C in acute kidney injury [D]. Nanjing Medical University, 2017. [Google Scholar]
- 37.Yu CJ. Changes in Blood Cys-C and Urine Kim-1 Concentrations After Cardiopulmonary Bypass Surgery and Their Diagnostic Value for AKI[D]. Southern Medical University.2013. [Google Scholar]
- 38.Zheng XF, Study and Validation of Cystatin C in Predicting Postoperative Acute Kidney Injury and Outcomes in Surgical Patients.Shantou University[D]. 2021. [Google Scholar]
- 39.Ronco C., Bellomo R., and Kellum J.A., Acute kidney injury. Lancet, 2019. 394(10212): p. 1949–1964. doi: 10.1016/S0140-6736(19)32563-2 [DOI] [PubMed] [Google Scholar]
- 40.Wen Y. and Parikh C.R., Current concepts and advances in biomarkers of acute kidney injury. Crit Rev Clin Lab Sci, 2021. 58(5): p. 354–368. doi: 10.1080/10408363.2021.1879000 [DOI] [PubMed] [Google Scholar]
- 41.Liu C., et al., Age- and sex-specific reference intervals for the serum cystatin C/creatinine ratio in healthy children (0–18 years old). J Int Med Res, 2019. 47(7): p. 3151–3159. doi: 10.1177/0300060519855575 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Hapca S., et al., The Relationship between AKI and CKD in Patients with Type 2 Diabetes: An Observational Cohort Study. J Am Soc Nephrol, 2021. 32(1): p. 138–150. doi: 10.1681/ASN.2020030323 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Gui Y., et al., Acute kidney injury in diabetes mellitus: Epidemiology, diagnostic, and therapeutic concepts. Faseb j, 2023. 37(4): p. e22884. doi: 10.1096/fj.202201340RR [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Karakasis P., et al., Effect of tirzepatide on albuminuria levels and renal function in patients with type 2 diabetes mellitus: A systematic review and multilevel meta-analysis. Diabetes Obes Metab, 2024. 26(3): p. 1090–1104. doi: 10.1111/dom.15410 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
(DOCX)
Data Availability Statement
All relevant data are within the paper and its Supporting Information files. Please see Tables 1 and 3, as well as the corresponding references (15-38).