Citations in English language from the national Library of Medicine's Medline database from 1989 to June 20, 2023 were evaluated by two authors (F.F., and C.M.A.). Initial testing for HCV was made in 1989 and data on HCV status are therefore not available for the time before 1989. The current research was performed by four Medline databases engines (Embase, Grateful Med, Ovid, and PubMed), and was limited to human studies.

The following algorithm in medical subject heading and in free text words was applied: (‘HCV’ or ‘HCV antibody positive serologic status’ or ‘HCV viremia’ or ‘Hepatitis C’ or ‘Hepatitis C Virus Infection’ or ‘HCV RNA’) AND (‘Albuminuria’ or ‘Macroalbuminuria’ or ‘Microalbuminuria’ or ‘Proteinuria’ or ‘Albumin/creatinine ratio’), AND (‘Incidence’ or ‘Prevalence’).

We performed an additional search with electronic searches of the Cochrane Library; manual searches of selected specialty journals were made to identify all pertinent literature. We also searched reference lists from qualitative topic reviews and published clinical studies. It was previously demonstrated that a Medline search alone might not be sensitive enough.6 Data on study design, study period, patient characteristics, HCV prevalence, antiviral therapy towards HCV, and kidney disease outcomes were abstracted. Authors of selected papers were contacted to obtain missing data and only information from patients with known HCV status were included in the meta-analysis. We achieved consensus for all data. We deleted duplicate reports for the same patients, which included contact with authors when necessary. We pre-specified eligibility and exclusion criteria. Our search was limited to human studies that were published in the English literature.

We retrieved studies if they met numerous inclusion criteria as follows:

• •

  They presented original data from cohort and longitudinal studies;

• •

  The outcome of interest was clearly defined as frequency of detectable proteinuria in the adult general population according to HCV status;

• •

  The studies provided adjusted risk estimates and their confidence intervals. Both case–control and cohort studies were considered as eligible for inclusion in the analysis.

If data on the same population were duplicated in more than one report, we enrolled the most recent study in the analysis. Information on HCV status was identified at the time of enrolment. We included reports where the diagnosis of HCV infection was conducted by testing for anti-HCV antibody in serum and/or detection of HCV RNA by nucleic acid testing. Surveys based on administrative codes (ICD-9) were also addressed.

We have excluded studies if they reported inadequate information or data on the link between proteinuria and HCV positive status. We have excluded unpublished studies and reports that were published in abstract form only. Letters and reviews (narrative or systematic reviews, meta-analyses) were not considered.

The quality of the 23 studies was evaluated using a scale from the ‘Newcastle–Ottawa Scale (NOS)’.7 The Newcastle–Ottawa Scale is a scoring system that assesses all aspects of an observational epidemiologic survey from a methodological point of view. When a study included relevant information that could be associated with the NOS, one point was added. Seven items in cross-sectional studies and eight items in cohort and case–control studies that could be related to the NOS were retrieved. Therefore, cross-sectional studies assigned 8–10, 6–7, 4–5, or 0–3 points (stars) were assessed as very good, good, satisfactory or unsatisfactory studies, respectively. Similarly, cohort and case–control studies with 7–9, 5–6, 4 and 0–3 points (stars) were retrieved as very good, good, satisfactory or unsatisfactory studies, respectively. We carried out subgroup analyses based on those reports provided with very good quality. Data extraction and quality scoring were performed independently by two reviewers (F.F., and F.M.D.) and the results were merged by consensus. The complete protocol for quality scoring is available on-line.

We made separate meta-analyses according to the outcome. The current meta-analysis included cross-sectional studies and longitudinal studies addressing the prevalence and incidence of proteinuria, respectively. The primary end-point was to provide adjusted estimates of the risk (and 95% CIs) of incidence (or prevalence) of proteinuria according to HCV infection. Multivariate analysis was conducted in each study to evaluate the independent impact of HCV positive status on the frequency of proteinuria after adjustment for potential confounders (covariates) (e.g., age, gender, diabetes mellitus, and others). Cox proportional hazard regression analysis and logistic regression analysis were carried out in longitudinal and cross-sectional studies, respectively.

An additional outcome was the occurrence of albuminuria according to HCV infection in studies without adjustment for covariates (univariate analysis only).

We weighted study-specific log hazard ratios (relative risks or odds ratios for cohort studies) by the inverse of their variances to obtain a pooled effect estimate and its 95% confidence intervals. We used the assumption that odds ratios provide a reasonable approximation of the RR in case control studies and in cohort studies in which the outcome occurs unfrequently (around 10%) in the unexposed population. In addition, we assumed that hazards can be thought as incidence rates (dichotomous, non-time-varying exposures) and thus the HRs can be roughly interpreted as incidence rate ratios. We combined cohort studies and case–control studies in the absence of statistical heterogeneity. For each study, we adopted the estimate of the effect measure that was adjusted for the largest number of confounders. When results from the fixed and random effects models were different, we presented the second as it represents a more conservative approach. We adopted the random-effects model, as reported by DerSimonian and Laird,8 and the heterogeneity was evaluated by Ri (i.e., the proportion of total variance due to between-study variance). Heterogeneity was considered consistent if Ri was ≥0.75. The origin of the heterogeneity was addressed adopting stratified analysis: we identified subgroups of studies defined by study characteristics such as country of origin, reference year, or size. Sensitivity analysis adopting a fixed-effect model was also conducted to evaluate the consistency of results. We assessed publication bias graphically by using a funnel plot as well as quantitatively with Egger's regression. Multivariate meta-regression analysis was performed with the aim to observe the impact of continuous covariates on the outcome (prevalence or incidence of proteinuria among HCV-infected patients). Statistical analysis was performed with the software HEpiMA (version 2.1.3)9 and Comprehensive Meta-analysis (Biostat Inc., USA, 2005). A 5% significance was adopted for the alpha risk. Every estimate was given with its 95% CI.

As listed in Fig. 1, we identified 452 articles and 68 were evaluated for eligibility. The list of the 68 full-text papers is reported available on request. Twenty-three studies met our inclusion criteria and were published in 22 papers10–31 and carried out in 3 continents (Europe, Asia and America); a total of 198,675 unique patients were retrieved. Thus, one study contributed data on more than one kidney disease outcome; however, the cohort was considered once in our meta-analysis. There was a 100% concordance between reviewers with regard to final inclusion and exclusion of studies reviewed based on the predefined inclusion and exclusion criteria.

Fig. 1.

Flow diagram of study selection.

(0,21MB).

Information on HCV status was collected at the time of study inclusion. We pooled studies where the diagnosis of HCV infection and chronic kidney disease were done by administrative codes (ICD-9-CM codes).

Tables 1 and 2 report some salient demographic and clinical characteristics of subjects enrolled in the included studies. The mean age of subject cohorts ranged from 38.5 to 60.8 years. The gender distribution ranged from 0 to 100% male. The average follow-up was between 15 and 132 months (longitudinal studies) and the pooled prevalence of HCV positive individuals was 0.19 (95% CI, 0.17; 0.21). Nine studies were from the USA, seven from Asia, five from Europe, and one from southern America. The quality of study scores ranged between 5 and 7 (longitudinal studies) and 6 and 9 (cross-sectional studies) (data not shown). The great majority of studies made diagnosis of HCV infection by detecting anti-HCV antibody in serum; one study assessed the presence of HCV antigen14 or HCV RNA24 in serum. Information on diagnosis of HCV was missed in one survey.23

Sixteen studies (n=163,797 unique patients; n=9854 HCV-positive and n=153,943 HCV-negative patients) evaluated the prevalence of proteinuria among HCV-infected patients.10–25 The pooled prevalence of proteinuria was 0.11 (95% CI, 0.09–0.14) (Table 3). We found a significant relationship between positive HCV status and increased prevalence of proteinuria, aOR 1.47 (95% CI, 1.3; 1.66) (P<0.001). The funnel plot concerning the publication bias is reported in Fig. 2; the Egger test reported significant publication bias (P=0.003). Tests for homogeneity of the aOR across the studies gave a Q value (by χ2 test) of 27.3 (P=0.026). Fig. 3 reports the aOR and 95% confidence intervals for each cross-sectional study.

Fig. 2.

Funnel plot of precision by log odds ratio (n=16 cross-sectional studies; n=163,979 unique patients) (outcome: prevalence of proteinuria).

(0,12MB).

Fig. 3.

aOR and 95% confidence intervals for each study (n=16 cross-sectional studies; n=163,979 unique patients) (outcome: prevalence of proteinuria). aOR of proteinuria associated with exposure to HCV: 1.47 (95% CI, 1.3; 1.66) (P<0.001) Ri=0.61. Q value by χ2 test 27.3 (P=0.026) (asymptotic tests).

(0,09MB).

Study 1 Szczech L, et al. 2002; study 2 Liangpunsakul S and Chalasani N, 2015; study 3 Tsui J, et al. 2006; study 4 Huang J, et al. 2006; study 5 Ishizaka N, et al. 2008; study 6 Yanik E, et al. 2010; study 7 Derbala M, et al. 2010; study 8 Lee J, et al. 2010; study 9 Yanagisawa N, et al. 2011; study 10 Aoufi Rabih S, et al. 2012; study 11 Reynes J, et al. 2013; study 12 Zeng Q, et al. 2014; study 13 Kurbanova N, et al. 2015; study 14 Antonello V, et al. 2015; study 15 Lucas G, et al. 2016; study 16 Zeder A, et al. 2016.

Seven longitudinal studies (n=35,170 unique patients; n=5556 anti-HCV positive and n=29,614 anti-HCV negative) gave information on the incidence of proteinuria. The pooled incidence of proteinuria was 0.14 (95% CI, 0.04; 0.23). The summary estimate of adjusted HR with HCV across the surveys, 1.79 (95%CI, 1.17; 2.74) (P<0.001).24,26–31 Significant publication bias occurred (P<0.04), according to the Egger test. Tests for homogeneity of the aHR across the studies (n=7) gave a Q value of 27.82 by χ2 test (P=0.0001).

Stratified analyses were carried out and did not show consistent differences in pooled aOR across various subgroups (Tables 4 and 5) but the homogeneity assumption was rejected in numerous subsets (Tables 4 and 5). As listed in Table 6, we carried out multivariate meta-regression analysis and found a significant link between percentage of HCV-infected males and adjusted OR (cross-sectional studies). The limited number of studies with longitudinal design (n=7) and the incomplete information on many covariates precluded the conduction of meta-regression in this setting. We collected incomplete information on many clinical or biochemical parameters and this hampered the completion of many comparisons.

We recorded studies (n=6; n=72,551 unique patients) addressing the occurrence of proteinuria according to HCV status (no adjustment for covariates made).32–37 There was an increased frequency of proteinuria among HCV-exposed individuals and this reached statistical significance, pooled OR 1.54 (95% CI, 1.08; 2.19). The homogeneity assumption was not rejected, Ri=0.88, Q value of 36.1 by χ2 test (P=0.001).