The rationale, objective and search strategy of this systematic review were registered in the International Prospective Registers of Systematic Reviews (PROSPERO) under the registration number CRD42021271501.

A systematic review of observational studies with uEGF determination was performed. Studies were searched in Medline, Google Scholar, Science Direct, and EBSCO up to August 2, 2021. Inclusion criteria: the presence of kidney disease, presence of urinary cancer, presence of kidney disease risk factors; in cross-sectional, case-control, or cohort studies, with uEGF determined by enzyme immunoassay (ELISA or EIA) or multiplex magnetic bead-based assay; language, and with availability of the full text. Exclusion criteria: reviews, clinical trials, pre-clinical studies, letters to the editor, or conference posters were excluded. Studies that did not report uEGF levels as means or medians were also excluded (for example, studies that showed data by tertile only).

The search strategy structure adopted was based on a PICO-style approach: Problem: human kidney disease; Intervention or prognostic factor: uEGF; Comparison: healthy or without kidney disease risk factors participants; Outcome: presence, absence, or progression of kidney disease.

We consider kidney disease as any primary or secondary nephropathy, ureteropelvic dysfunction or hydronephrosis, renal cancer, AKI, or CKD. Different clinical settings included ICU, hospitalised, or outpatient milieu.

The electronic search strategy for Medline was carried out with the following terms:

((((((((((EGF [Title/Abstract]) OR EPIDERMAL GROWTH FACTOR [Title/Abstract]) OR EPITHELIAL GROWTH FACTOR [Title/Abstract])) OR EPIDERMAL GROWTH FACTOR [mesh]))) OR (EGF [MeSH Major Topic]) AND (((((KIDNEY [Title]) OR NEPHROPATHY [Title/Abstract]) OR RENAL [Title])) OR ((((NEPHROPATHY [MeSH Major Topic]) AND RENAL DISEASE [MeSH Major Topic]) AND KIDNEY DISEASE [MeSH Major Topic]))))) NOT REVIEW[Publication Type] AND (humans[Filter])) NOT (cells[Title]), and the search strategy for other databases is presented in the supplementary material. Reference lists, similar articles or those cited by another article of identified articles, as well as other review studies, were also reviewed manually to identify additional articles. The MOOSE and PRISMA guidelines for reporting systematic reviews were followed4,5 and quality assessment was performed to assess potential risk of bias for each included study according to the NIH/NHLBI Quality Assessment Tools6; depending on the methodological design as cross sectional/cohort, or cases and controls, the respective NIH/NHLBI Quality Assessment Tool was applied; and, the methodological design as indicated by the authors was considered, whereas if it was not specified, or was only mentioned as a prospective study, the methodological design corresponding to the methodology described in the study was identified. The quality of the studies was classified as good, fair, or poor (see supplementary material).

All reviewers are researchers or students from the health area. One person extracted the data, and another person checked the extracted data. Disagreement was discussed and consensus was reached using a third opinion. Two reviewers independently assessed potential risk of bias and were blinded to each other. Disagreement was discussed and a consensus reached using a third opinion. Studies in languages other than English or Spanish were translated using an online Google translator.

Studies were grouped according to the authors’ declaration as CKD, AKI, or a specific nephropathy, the latter could present different degrees of renal function. An Excel datasheet was used for data extraction. The variables extracted included: age, sex, sample size, settings, biological material in which uEGF was measured (spot or 24-h urine), uEGF measurement technique, sample storage, uEGF values, and its units of measurement. Also, specificity, sensitivity, area under the ROC curve (AUC), and hazard ratio (HR) with 95% confidence interval (CI) were extracted whenever they were available.

The initial results of the bibliographic search identified 936 articles, from which 342 were eliminated because they were duplicates, 494 were excluded based on title or abstract review or not retrieved and 19 for using radioimmuno-assay. The main reason for excluding articles after reading the full text was that they did not evaluate uEGF levels. After reading the full articles, 53 studies were ultimately included.7–59 The flowchart for the selection process was according to PRISMA guidelines, and Fig. 1 shows a flowchart of the study's selection process. The characteristics of studies included are presented in Supplementary Tables 1–4.

Fig. 1.

Flowchart of the study selection process.

(0,57MB).

Supplementary Table 5 shows the evaluation of the quality of the selected articles. The most frequently found risks of bias were the lack of justification of the sample size power description or variance and effect estimates provided, the lack of a specified and defined study population, and the lack of measurement and analysis of key potential confounding variables. None of the studies reported whether there was blinding to the exposure status of participants of outcome assessors. From the included studies in this review, 30 (56.6%) were identified as “good” quality, 22 (41.5%) were classified as “intermediate”, and 1 (1.8%) were classified as “poor” quality.

Several studies reported more than one type of kidney disease: the most frequently studied diseases were AKI (20.7%), CKD (18.8%), diabetic nephropathy (9.4%), and reflux or obstructed nephropathy (NPT) (7.5%); while other kidney diseases were: glomerulonephritis, IgA NPT, post-transplantation renal tumour, polycystic kidney disease (PKD), congenital anomalies, lupus nephritis, carcinoma of the bladder, renal amyloidosis, Henoch-Schönlein purpura nephritis, renal pelvic or calyceal stone, Alport syndrome, Wilms tumour, human immunodeficiency virus (HIV) NPT, and unspecified origin NPT.

The methodological design reported by the authors or identified by the reviewers were, according to inclusion criteria, 34 studies with cohort design, 15 cross-sectional, and 4 with case-control design. The uEGF assays used in the studies were Enzyme Immunoassay/Enzyme-Linked Immunosorbent Assay (ELISA), and multiplex assay. The oldest studies were those using radioimmunoassay (RIA); most of them with a low-quality assessment, so they were not included in the review, while the most recent studies used both ELISA and multiplex assay (a type of immunoassay to simultaneously measure multiple analytes in a single sample). Regarding the units reported for uEGF levels, 39 studies reported uEGF adjusted to urinary creatinine, 9 studies reported uEGF in weight units over volume units (ej, pg/mL), 3 studies presented the levels in weight units over 24h (ej. pg/24-h), one study reported as a rate (mg/mL/min), and one study did not specify units.

The uEGF levels were analysed in 39 studies in both primary and secondary kidney diseases as well as in CKD; they can be seen in Table 1. According to the age of the patients studied, 13 studies were performed on children and adolescents, and 26 on adults and older adults, one of them did not mention the age of the participants.7 Among the studies that reported Cr-adjusted uEGF values, the lowest values in patients with kidney disease were found in adult patients by Stangou et al., in IgA NPT and in pauci-immune focal segmental necrotising glomerulonephritis12,13; however, the controls of these patients presented with the lowest levels of uEGF reported where, in both cases, values equal to or less than 0.00014ng/mg Cr were reported; while the highest values in patients with kidney disease were 80ng/mg Cr in paediatric patients with obstructive nephropathy,33 and the highest values for controls were 120.6ng/mg Cr in adults.19

Table 2 shows the studies that analysed and reported uEGF levels in children or adults patients with AKI. Values reported in these cases were from 1.7ng/mg Cr in adults45 to 18.8ng/mg Cr in children hospitalised in the ICU.47 In the 4 studies regarding AKI, all of them reported lower levels of uEGF in cases versus healthy or exposed controls.

Malignant neoplasms of the kidney or bladder were evaluated only in one of the included studies48 where paediatric Wilms tumour survivors with eGFR<90mL/min/1.73m2, uEGF levels were lower.

The uEGF was evaluated in 9 studies carried out in neonates, in all of them reported the diagnostic accuracy analysis values and 6 also showed the specific values of uEGF (see Table 3); of which, the lowest values were those reported in neonates of term with, and without, AKI by Askenazie et al.52; yet when adjusting the values with urinary Cr levels, the values reported in neonates with AKI and with ureteropelvic obstruction were similar.

In all cases, uEGF values were lower in cases with kidney disease versus healthy controls or those without kidney disease.

Of the studies included, only 7 in adults, 1 in children and 6 in neonates reported a cut-off value: 13 of them showed the AUC of the respective cut-off value, and 11 reported sensitivity and specificity (see Table 4). In adults, the lowest cut-off values in ng/mg Cr or ng/mL were observed in CKD and in antibody-mediated kidney allograft rejection,23,29 while the highest values were described in primary glomerulonephritis.19 The highest AUC values were observed with cut-off values of 10.8ng/mg Cr to identify primary glomerulonephritis, and 5.3ng/mg Cr for lupus nephritis: these cut-off values also presented the highest sensitivities.17,30

In neonates, the lowest cut-off value (1.75pg/mL) was observed in patients with AKI treated with hypothermia for hypoxic ischaemic encephalopathy,56 and when the uEGF was adjusted for the level of urinary Cr, the cut-off values in neonates with AKI were higher, reaching up to 45ng/mg Cr; where this last value was the one that reported the highest AUC as well as sensitivity and specificity in neonates.50

No cut-off points were identified for studies related to AKI, or kidney or bladder cancer, in children or adults.