Factors Correlating to the Development of Hepatitis C Virus Infection in Hemodialysis Patients—Findings Mainly from Asiatic Populations: A Systematic Review and Meta-Analysis

Hemodialysis is an effective replacement therapy for chronic renal failure patients. In recent decades, the number of hemodialysis patients has grown rapidly and some measures for preventing blood-borne diseases have been implemented, but hepatitis C virus (HCV) infection remains a significant problem. The meta-analysis published in 2009 on HCV infection-related factors was based on localized study objects, and some additional studies have been published since 2009; however, the contribution of these factors remains under dispute. Our study pooled the odds ratios (ORs) or mean standard deviations (MDs) with 95% confidence intervals (CIs) and analyzed sensitivity using Review Manager 5.1 software (5.1 version Copenhagen: The Nordic Cochrane Centre; 2011) by searching data in the PubMed, Elsevier, Springer, Wiley, and EBSCO databases. Spearman correlation analysis was performed using the SPSS package. In our meta-analysis, 1715 HCV-infected hemodialysis patients and 7093 non-HCV-infected hemodialysis patients from 44 studies were analyzed. The pooled ORs with 95% CIs were: histories of blood transfusion, 4.30 (3.11, 5.96); weekly hemodialysis times > 2, 6.00 (3.25, 11.06); kidney transplantation, 5.80 (3.95, 8.52); hemodialysis units > 2, 6.90 (2.42, 19.68); shared hemodialysis devices, 5.00 (2.35, 10.65); and drug addiction, 4.73 (1.54, 14.47). The pooled MDs with 95% CIs were duration of hemodialysis (months) 27.48 (21.67, 33.30). There was a positive correlation between duration of hemodialysis and the HCV infection rate (p < 0.01). Hemodialysis patients, especially from Asia, with shared hemodialysis devices, hemodialysis units > 2, blood transfusion, kidney transplantation, and drug addiction were at increased risk of HCV infection. The HCV infection rate increased with the duration of hemodialysis. High-risk hemodialysis patients should be monitored and receive timely screening.


Introduction
Hemodialysis is an effective replacement therapy for chronic renal failure patients that can increase survival times [1]. In recent decades, the number of patients administered hemodialysis has grown

Inclusion and Exclusion Criteria
In this study, the eligibility criteria for the inclusion of literature in the meta-analysis were as follows: (1) the literature is the original research; (2) the literature was an observational study with specific temporal and geographic characteristics; (3) the literature was published with the full text available; (4) all cases and controls were hemodialysis patents and the source of samples was clearly stated; (5) hepatitis C was diagnosed according to the national diagnostic criteria that existed at that time [64] and possible risk factors were reported; and (6) the literature was published in Chinese or English.
Literature was excluded from the meta-analysis when (1) the data reported could not be used to calculate the odds ratio (OR) or mean standard deviation (MD) with 95% confidence interval (CI) for the main variable; (2) the literature duplicated the same research; (3) the literature used the same research objects; and (4) the literature was deemed to be of poor quality literature (based on Ebrahim et al.'s declaration, the number of items satisfied in the corresponding research type declaration was less than half of total items) [65,66].

Data Extraction
We used a pre-made form for data extraction, and then two trained reviewers assessed the literature one by one and completed the form. The following data were extracted from the qualified studies: first author, year of the study, location, the number of hemodialysis patients in the HCV-infected group and the non-HCV-infected group, sample size, male to female ratio, and age distribution for HCV infection development among hemodialysis patents.
Discrepancies between the assessment results obtained by the two reviewers were resolved by discussion and checking the original documents.

Sensitivity Analysis
The studies with the widest 95% CI for the OR or MD were omitted from the subgroup analysis for this factor, and the remaining studies were pooled and pooled MD CI or pooled OR CI values with 95% CIs were obtained for this study factor, and then this pooled MD CI or pooled OR CI was compared with the total pooled OR or pooled MD before omitting this study factor. The studies with the maximum weight were omitted from the subgroup analysis, and then pooled, and the pooled MD weight or OR weight values with 95% CI for this study factor were obtained and then this pooled MD weight or pooled OR weight was compared with the total pooled MD or pooled OR before omitting this study factor.
In this meta-analysis, subgroup analyses were used to determine the associations between different study factors and HCV infection, a sensitivity analysis was used to examine the reliability of the associations, and a funnel plot was used to examine publication bias. According to the manufacturer's instructions, the normal ranges of values for serum alanine aminotransferase (ALT) are about 5-40 units per liter of serum. Based on the standard procedure reported in the instructions, HCV infection was confirmed when a serum sample tested positive for HCV antibodies.

Statistical Analysis
The OR or MD with 95% CI were taken as the main indicators in this meta-analysis. Review Manager 5.1 software (5.1 version Copenhagen: The Nordic Cochrane Centre; 2011) was used to analyze the fixed-effect model without heterogeneity or the random-effect model with heterogeneity, after the heterogeneity test. The heterogeneity among different studies for study factors was evaluated by Cochran's chi-square test with a significance level α = 0.1 and I 2 statistics. The OR or MD was not pooled when its number for the study factor was less than 4. In the meta-analysis, I 2 statistics, ranging from 0% to 100%, were used to assess the levels of heterogeneity; values of 0%, 25%, 50%, 75%, and 100% were taken as no, low, medium, high, and significant heterogeneity, respectively [67]. In this statistics. The OR or MD was not pooled when its number for the study factor was less than 4. In the meta-analysis, I 2 statistics, ranging from 0% to 100%, were used to assess the levels of heterogeneity; values of 0%, 25%, 50%, 75%, and 100% were taken as no, low, medium, high, and significant heterogeneity, respectively [67]. In this meta-analysis, I 2 ≤ 50% was accepted. Correlation analysis was performed using SPSS version 16 software (SPSS Inc., Chicago, IL, USA). Spearman correlation was used for ranked data, with α = 0.05 considered to indicate statistical significance.

Literature Search
In this meta-analysis, a total of 44 research articles were included, and a flow chart of the literature selection process is shown in Figure 1. Of the literature selected, 44 studies included 1715 HCV-infected hemodialysis patients and 7093 non-HCV-infected hemodialysis patients. The study characteristics, region, study type, number of HCV-infected hemodialysis patients and non-HCV-infected hemodialysis patients, study factors, sample size, male/female ratio, and mean participant age (years) are shown in Table 1. Of the literature selected, 44 studies included 1715 HCV-infected hemodialysis patients and 7093 non-HCV-infected hemodialysis patients. The study characteristics, region, study type, number of HCV-infected hemodialysis patients and non-HCV-infected hemodialysis patients, study factors, sample size, male/female ratio, and mean participant age (years) are shown in Table 1. The 11 study factors used to pool OR or MD with 95% CI were as follows: gender (15 studies, 867 cases, 3874 controls), age (16 studies, 953 cases, 4770 controls); a history of blood transfusion (28 studies, 820 cases, 3663 controls); weekly hemodialysis times > 2 (5 studies, 103 cases, 344 controls); a history of kidney transplantation (8 studies, 234 cases, 1507 controls); hemodialysis units > 2 (6 studies, 127 cases, 389 controls); shared hemodialysis devices (6 studies, 275 cases, 1358 controls); serum alanine aminotransferase (ALT) levels (abnormal) (6 studies, 280 cases, 1344 controls); drug addiction (4 studies, 86 cases, 639 controls); a history of surgery (7 studies, 161 cases, 1123 controls); and duration of hemodialysis (months) (28 studies, 940 cases, 4044 controls).

Results of Pooled ORs or MDs
In this meta-analysis, the pooled ORs and their 95% CIs for study factors were as follows: histories of blood transfusion, 4.30 (3.11, 5.

Results of Pooled ORs or MDs
In this meta-analysis, the pooled ORs and their 95% CIs for study factors were as follows: histories of blood transfusion, 4.30 (3.11, 5.

Results of Heterogeneity Evaluation
A heterogeneity test for pooled ORs with 95% CIs showed that variations among ORs for study factors including histories of blood transfusion, hemodialysis units > 2, shared hemodialysis devices, and serum ALT levels were statistically significant (p < 0.10). The effects of these factors were then pooled using the random-effect model, whereas weekly hemodialysis times, kidney transplantation, drug addiction, a history of surgery, and gender were pooled using the fixed-effect model (p > 0.10). A heterogeneity test for pooled MDs with 95% CIs showed that the variation among studies for the duration of hemodialysis (months) was statistically significant (p < 0.10). The effects were then pooled using the random-effect model. These results are detailed in Figures 2-5.

Results of Heterogeneity Evaluation
A heterogeneity test for pooled ORs with 95% CIs showed that variations among ORs for study factors including histories of blood transfusion, hemodialysis units > 2, shared hemodialysis devices, and serum ALT levels were statistically significant (p < 0.10). The effects of these factors were then pooled using the random-effect model, whereas weekly hemodialysis times, kidney transplantation, drug addiction, a history of surgery, and gender were pooled using the fixed-effect model (p > 0.10). A heterogeneity test for pooled MDs with 95% CIs showed that the variation among studies for the duration of hemodialysis (months) was statistically significant (p < 0.10). The effects were then pooled using the random-effect model. These results are detailed in Figures 2-5.

Publication Bias
In the meta-analysis, a funnel plot of the articles including the duration of hemodialysis was symmetrical, with the axis of symmetry (MD = 0) being to the right of center, as detailed in Figure 6.

Sensitivity Analysis
In view of the reliability of the pooled ORs or MDs using the random-effect model for terms including histories of blood transfusion, shared hemodialysis devices, hemodialysis units > 2, serum ALT levels, and duration of hemodialysis (months), we omitted studies with the widest 95% CIs for the ORs and MD values, respectively, and pooled and acquired ORCI and MDCI values with the 95% CIs, and these pooled values were close to the respective pooled OR and MD values with 95% CIs, as detailed in Table 2.
In view of the reliability of pooled ORs using the random-effect model for terms including histories of blood transfusion, shared hemodialysis devices, hemodialysis units > 2, serum ALT levels, and duration of hemodialysis (months), we omitted studies with the highest weights, and pooled and acquired ORweight or MDweight values with 95% CIs, and these pooled values were close to the respective pooled OR or MD values with 95% CIs, as detailed in Table 2.

Sensitivity Analysis
In view of the reliability of the pooled ORs or MDs using the random-effect model for terms including histories of blood transfusion, shared hemodialysis devices, hemodialysis units > 2, serum ALT levels, and duration of hemodialysis (months), we omitted studies with the widest 95% CIs for the ORs and MD values, respectively, and pooled and acquired OR CI and MD CI values with the 95% CIs, and these pooled values were close to the respective pooled OR and MD values with 95% CIs, as detailed in Table 2. In view of the reliability of pooled ORs using the random-effect model for terms including histories of blood transfusion, shared hemodialysis devices, hemodialysis units > 2, serum ALT levels, and duration of hemodialysis (months), we omitted studies with the highest weights, and pooled and acquired OR weight or MD weight values with 95% CIs, and these pooled values were close to the respective pooled OR or MD values with 95% CIs, as detailed in Table 2.

Discussion
This study showed that, for hemodialysis patients, the rate of HCV infection increased with the duration of hemodialysis treatment. This meta-analysis also found that hemodialysis patients with a duration of hemodialysis treatment >5 years and/or histories of blood transfusion and/or shared hemodialysis devices and/or hemodialysis units >2 and/or weekly hemodialysis times >2 and/or kidney transplantation and/or histories of surgery and/or drug addiction were at increased risk of developing HCV infection, whereas the age and gender of hemodialysis patients did not affect the risk of developing HCV infection.
Our study analyzed the rate of HCV infection among groups of patients with a duration of hemodialysis treatment of 1-2 years, 2-3 years, 3-5 years, 5-10 years, and >10 years, and the study found that the longer the duration of hemodialysis treatment, the higher the rate of HCV infection; this result was consistent with the results reported in the meta-analysis by Sun et al. in 2009 [8].
More specifically, our study showed that patients with a duration of hemodialysis treatment >5 years were at increased risk of developing HCV infection, whereas patients with a duration of hemodialysis treatment <5 years did not have an increased risk of developing HCV infection. This result was not consistent with the findings of the 2009 meta-analysis, which reported that patients with a duration of hemodialysis treatment >1 year were at increased risk of developing HCV infection [8]. These differences may reflect the implementation of effective management measures imposed by relevant healthcare organizations in recent years.
In our meta-analysis, the result of a quantitative analysis (Figure 4) also showed that the duration of hemodialysis for HCV-infected patients was 27.48 months longer than the duration of hemodialysis for non-HCV-infected patients. This finding was longer than the 15.41 months reported in the 2009 meta-analysis by Sun et al. [8], and this difference may also reflect the implementation of the above-mentioned effective management measures in recent years. Moreover, this result of the quantitative analysis was also consistent with those of qualitative analysis in this study (i.e., patients with a duration of hemodialysis treatment >5 years did show an increased risk of developing HCV infection, whereas patients with a duration of hemodialysis treatment < 5 years were not at higher risk).
In general, exposure to HCV-contaminated medical equipment or goods can increase the risk of HCV infection, and during the process of hemodialysis, patients have many possible opportunities for exposure to HCV-contaminated equipment or goods [68,69].
The findings of this meta-analysis showed that hemodialysis patients with a history of shared hemodialysis devices and/or hemodialysis units >2 and/or weekly hemodialysis times >2 and/or a duration of hemodialysis >5 years were at increased risk of HCV infection, and this may be related to the fact that these hemodialysis patients had more opportunities to be exposed to HCV-contaminated medical equipment, HCV-contaminated goods, or the HCV-contaminated hands of medical personnel, potentially leading to nosocomial infection.
A study by Alfurayh et al. confirmed the existence of nosocomial transmission in hemodialysis centers by sequence analysis [70]. Moreover, the findings of this meta-analysis showed that hemodialysis patients with a history of drug addiction were at increased risk of HCV infection and this may be related to the fact that these hemodialysis patients had shared HCV-contaminated needles and syringes, leading to cross infection. From what has been discussed above, we suggest that disposable goods, such as disposable dialysis dialyzers, disposable dialysis pipes, and so on, should be used to cut off cross infection during hemodialysis.
The findings of this meta-analysis also showed that kidney transplantation hemodialysis patients were at increased risk of HCV infection, and this may be related to the fact that these hemodialysis patients had taken immunosuppressants, which may have resulted in low lymphocyte activation following HCV infection [71]. Moreover, the findings of this meta-analysis also showed that hemodialysis patients with abnormal serum ALT levels were at increased risk of HCV infection, and this may be related to the chronological order of the development of abnormal elevated serum ALT levels, which could not be identified in the observational studies included, or the fact that these hemodialysis patients had disrupted normal liver structure and function, which may have resulted in low lymphocyte activation following HCV infection.
In general, ELISA was the routine method for screening blood donors for HCV infection, but molecular-based tests such as PCR are more sensitive diagnostic assays, and thus it is possible that some blood donors screened by traditional ELISA methods may have been HCV infectors [72][73][74]. This may explain our finding that hemodialysis patients with histories of blood transfusion were at higher risk of developing HCV infection. Thus, we suggest that blood donors and hemodialysis patient populations should be tested regularly with more sensitive PCR diagnostic assays.
Our meta-analysis also found that the age and gender of hemodialysis patients did not affect the risk of developing HCV infection, and this result was consistent with the findings of the 2009 meta-analysis [8].
The sensitivity analysis performed as part of our meta-analysis found that, after omitting studies with the widest 95% CIs for OR or MD values and studies with the maximum weight in subgroup analyses for the duration of hemodialysis treatment (months), histories of blood transfusion, shared hemodialysis devices, hemodialysis units >2, and abnormal serum ALT levels, the two overall effects were not reversed and the pooled OR or MD values were similar to those observed before omitting the studies. This revealed that the pooled ORs or MDs for these study factors were reliable and stable.
The limitations of this study were that only articles published in English or Chinese were included in the meta-analysis. In addition, even though the ORs or MDs of the six factors were pooled using a random-effect method, study heterogeneity may have influenced the findings to some extent. Furthermore, some study factors, for example, the degree of deterioration and the socioeconomic status of patients, were not available to be pooled. Lastly, studies included were carried out in the following countries: Iran, Australia, Egypt, and China, and this aspect limits the generalizability of conclusions.

Conclusions
It can be concluded that, for hemodialysis patients, the rate of HCV infection increases with the duration of hemodialysis treatment, and that hemodialysis patients, especially from Asia, with histories of blood transfusion and/or weekly hemodialysis times >2 and/or shared hemodialysis devices and/or hemodialysis units >2 and/or kidney transplantation and/or drug addiction were at increased risk of developing HCV infection. High-risk hemodialysis patients should be closely monitored and receive timely screening and therapeutic intervention to reduce the risk of HCV nosocomial infection.

Conflicts of Interest:
The authors declare that they have no conflict of interests.

List of Abbreviations
The following abbreviations are used in this manuscript: