Zinc is essential for the maintenance of normal cellular structure and functions. Zinc dyshomeostasis can lead to many diseases, such as cardiovascular disease. However, there are conflicting reports on the relationship between serum zinc levels and heart failure (HF). The purpose of the present study is to explore the relationship between serum zinc levels and HF by using a meta-analysis approach. PubMed, Web of Science, and OVID databases were searched for reports on the association between serum zinc levels and HF until June 2016. 12 reports with 1453 subjects from 27 case-control studies were chosen for the meta-analysis. Overall, the pooled analysis indicated that patients with HF had lower zinc levels than the control subjects. Further subgroup analysis stratified by different geographic locations also showed that HF patients had lower zinc levels than the control subjects. In addition, subgroup analysis stratified by HF subgroups found that patients with idiopathic dilated cardiomyopathy (IDCM) had lower zinc levels than the control subjects, except for patients with ischemic cardiomyopathy (ICM). In conclusion, the results of the meta-analysis indicate that there is a significant association between low serum zinc levels and HF.
Zinc is an important trace element in the body. It plays a critical role in maintaining cellular structure and functions and is also involved in gene expression and cell growth and differentiation as a catalytic and structural cofactor [
Heart failure (HF) is one of the most frequent causes of death worldwide. Recently, there has been emerging evidence suggesting that micronutrient dyshomeostasis is associated with HF [
A systematic literature search of PubMed, Web of Science, and OVID databases was done until June 2016 with the use of medical subject headings (MeSH) or free text words. The search keywords were “zinc” or “Zn” and “heart failure.” The references cited in the studies and in review articles were also reviewed to identify additional works that were not captured by the database search. Only published studies with full-text reports were included.
Three authors (Xuefang Yu, Lei Huang, and Jinyan Zhao) carried out the search independently. Titles and abstracts were screened for subject relevance, and studies that could not be definitely excluded based on the abstract information were selected for full-text screening. Two authors (Xuefang Yu and Lei Huang) independently selected eligible studies for possible inclusion in the meta-analysis. Any disagreement regarding study inclusion was resolved by discussion with Jinyan Zhao toward reaching a consensus.
The appropriateness of the studies was assessed. The criteria for inclusion in the analysis were (1) human study, (2) case-control or cohort study or randomized clinical trial, (3) focus on the association between serum zinc levels and HF, and (4) providing sufficient data on zinc levels in both HF patients and control subjects. The exclusion criteria were (1) obviously irrelevant study, (2) animal study, (3) review or case report, and (4) not providing data on zinc levels in either HF patients or control subjects.
All data were extracted independently by two reviewers (Bo Bian and Wei Yao) according to the inclusion and exclusion criteria. Specifically, the following data were extracted: authors’ names, year of publication, country, number of subjects, and data on serum zinc levels. Any inconsistencies or discrepancies in the extracted information were resolved by discussion between the two reviewers, with a third author (Zhuoqun Wang) providing input.
The quality of all the included studies was evaluated by using the Newcastle-Ottawa Scale (NOS). This assessment tool focused on three aspects: participant selection, comparability, and exposure. Studies that satisfied all the items on the scale were given nine stars. Two authors (Xianming Wu and Jingjing Huang) assessed the quality of studies independently.
The statistical analysis was carried out by using Stata 12, and statistical significance was set at
Subgroup analysis was done to determine associations between serum zinc levels and other relevant study characteristics, which may be possible sources of heterogeneity. Sensitivity analysis was carried out with one study removed at a time to assess whether the results could be affected markedly by a single study. Publication bias was measured by using Begg’s test and visualization of funnel plots.
Figure
Characteristics of subjects in included studies.
Studies | Country | Number of participants | Age (year) | Subgroups of HF | Score | ||
---|---|---|---|---|---|---|---|
HF | Controls | HF | Controls | ||||
Sullivan et al., 1979 | USA | 42 | 37 | NA | NA | ICM | 6 |
Atlihan et al., 1990 (1) | Turkey | 29 | 11 |
|
|
NA | 6 |
Atlihan et al., 1990 (2) | Turkey | 29 | 11 |
|
|
NA | 6 |
Oster, 1993 | Germany | 20 | 50 |
|
|
IDCM | 8 |
Golik et al., 1993 (1) | Israel | 9 | 20 |
|
|
NA | 8 |
Golik et al., 1993 (2) | Israel | 15 | 20 |
|
|
NA | 8 |
Cénac et al., 1996 | France | 35 | 36 | NA | NA | NA | 7 |
De Lorgeril et al., 2001 |
|
21 | 18 | 27–76 | 34–68 | CHF | 7 |
Topuzoglu et al., 2003 | Turkey | 54 | 20 | 18–75 | 21–73 | IDCM | 6 |
Kosar et al., 2006 (1) | Turkey | 54 | 30 |
|
|
CHF | 8 |
Kosar et al., 2006 (2) | Turkey | 26 | 30 |
|
|
IDCM | 8 |
Kosar et al., 2006 (3) | Turkey | 28 | 30 |
|
|
ICM | 8 |
Kosar et al., 2006 (4) | Turkey | 24 | 30 |
|
|
NA | 8 |
Kosar et al., 2006 (5) | Turkey | 20 | 30 |
|
|
NA | 8 |
Kosar et al., 2006 (6) | Turkey | 10 | 30 |
|
|
NA | 8 |
Salehifar et al., 2008 (1) | Iran | 18 | 27 |
|
|
IDCM | 8 |
Salehifar et al., 2008 (2) | Iran | 3 | 27 |
|
|
IDCM | 8 |
Salehifar et al., 2008 (3) | Iran | 4 | 27 |
|
|
IDCM | 8 |
Salehifar et al., 2008 (4) | Iran | 9 | 27 |
|
|
IDCM | 8 |
Salehifar et al., 2008 (5) | Iran | 2 | 27 |
|
|
IDCM | 8 |
Shokrzadeh et al., 2009 (1) | Iran | 30 | 27 |
|
|
ICM | 7 |
Shokrzadeh et al., 2009 (2) | Iran | 17 | 27 |
|
|
ICM | 7 |
Shokrzadeh et al., 2009 (3) | Iran | 10 | 27 |
|
|
ICM | 7 |
Ghaemian et al., 2011 (1) | Turkey | 38 | 40 |
|
|
NA | 8 |
Ghaemian et al., 2011 (2) | Turkey | 40 | 40 |
|
|
NA | 8 |
Alexanian et al., 2014 (1) | Greece | 81 | 21 |
|
|
NA | 8 |
Alexanian et al., 2014 (2) | Greece | 44 | 21 |
|
|
NA | 8 |
Forest plot of studies on zinc levels in HF patients versus control subjects. The combined SMD and 95% CIs were calculated by using a random-effects model.
There were 12 studies that assessed the association between serum zinc levels and HF. First, the heterogeneity among the included studies was assessed. The results showed a high statistical heterogeneity (
Summary of studies included in the analysis of serum zinc levels for subjects with HF versus control subjects.
Author | HF | Controls | Weight (%) | SMD (95% CI) | ||
---|---|---|---|---|---|---|
Number | Zn concentration (mean ± SD) | Number | Zn concentration (mean ± SD) | |||
Sullivan et al., 1979 | 42 | 0.74 ± 0.11 |
37 | 0.97 ± 0.13 |
4.09 |
|
Atlihan et al., 1990 (1) | 29 | 92.9 ± 18.9 |
11 | 107.5 ± 15.7 |
3.55 |
|
Atlihan et al., 1990 (2) | 29 | 107.4 ± 17.2 |
11 | 107.5 ± 15.7 |
3.62 |
|
Oster, 1993 | 20 | 745 ± 195 |
50 | 931 ± 178 |
4.06 |
|
Golik et al., 1993 (1) | 9 | 11.5 ± 1.53 |
20 | 16 ± 3.3 |
3.06 |
|
Golik et al., 1993 (2) | 15 | 13.2 ± 2.3 |
20 | 16 ± 3.3 |
3.57 |
|
Cénac et al., 1996 | 35 | 0.9 ± 0.21 |
36 | 1.17 ± 0.25 |
4.18 |
|
De Lorgeril et al., 2001 | 21 | 0.82 ± 0.12 mg/l | 18 | 0.9 ± 0.09 mg/l | 3.74 |
|
Topuzoglu et al., 2003 | 54 | 81.42 ± 15.4 |
20 | 92.51 ± 22.8 |
4.12 |
|
Kosar et al., 2006 (1) | 54 | 555 ± 104 |
30 | 620 ± 130 |
4.32 |
|
Kosar et al., 2006 (2) | 26 | 568 ± 116 |
30 | 620 ± 130 |
4.1 |
|
Kosar et al., 2006(3) | 28 | 542 ± 92 |
30 | 620 ± 130 |
4.1 |
|
Kosar et al., 2006 (4) | 24 | 565 ± 107 |
30 | 620 ± 130 |
4.06 |
|
Kosar et al., 2006 (5) | 20 | 543 ± 106 |
30 | 620 ± 130 |
3.96 |
|
Kosar et al., 2006 (6) | 10 | 553 ± 66 |
30 | 620 ± 130 |
3.52 |
|
Salehifar et al., 2008 (1) | 18 | 0.97 ± 0.25 mg/l | 27 | 1.12 ± 0.42 mg/l | 3.89 |
|
Salehifar et al., 2008 (2) | 3 | 1.3 ± 0.09 mg/l | 27 | 1.12 ± 0.42 mg/l | 2.31 | 0.444 ( |
Salehifar et al., 2008 (3) | 4 | 0.95 ± 0.26 mg/l | 27 | 1.12 ± 0.42 mg/l | 2.63 |
|
Salehifar et al., 2008 (4) | 9 | 0.87 ± 0.24 mg/l | 27 | 1.12 ± 0.42 mg/l | 3.39 |
|
Salehifar et al., 2008 (5) | 2 | 1.04 ± 0.06 mg/l | 27 | 1.12 ± 0.42 mg/l | 1.87 |
|
Shokrzadeh et al., 2009 (1) | 30 | 1.05 ± 0.28 mg/l | 27 | 1.12 ± 0.42 mg/l | 4.13 |
|
Salehifar et al., 2009 (2) | 17 | 1.27 ± 0.29 mg/l | 27 | 1.12 ± 0.42 mg/l | 3.86 | 0.399 ( |
Shokrzadeh et al., 2009 (3) | 10 | 0.95 ± 0.27 mg/l | 27 | 1.12 ± 0.42 mg/l | 3.5 |
|
Ghaemian et al., 2011 (1) | 38 | 24.7 ± 27.6 |
40 | 70.9 ± 21.6 |
4.09 |
|
Ghaemian et al., 2011 (2) | 40 | 23.2 ± 16.8 |
40 | 70.9 ± 21.6 |
3.94 |
|
Alexanian et al., 2014 (1) | 81 | 74.27 ± 15.87 |
21 | 87.9 ± 17.85 |
4.21 |
|
Alexanian et al., 2014 (2) | 44 | 81.51 ± 19.73 |
21 | 87.9 ± 17.85 |
4.12 |
|
The subgroup analysis showed that the geographic locations and etiologies of HF [idiopathic dilated cardiomyopathy (IDCM) and ischemic cardiomyopathy (ICM)] had some influence on the serum zinc levels in HF patients and control subjects.
Further subgroup analysis stratified by different geographic locations found that HF patients had lower zinc levels than the control subjects [Europe: SMD: −0.832 and 95% CI: −1.119, −0.545; Asia: SMD: −0.408 and 95% CI: −0.761, −0.055; America: SMD: −1.920 and 95% CI: −2.456, −1.384] (Figure
Subgroup analyses of zinc level and heart failure (HF).
Subgroup | Numberof studies | SMD (95% CI) | Test of SMD = 0 | Heterogeneity | ||
---|---|---|---|---|---|---|
|
|
|
|
|||
Geographical location | ||||||
Europe | 16 |
|
<0.001 | 5.68 | 76.1% | <0.001 |
Asia | 10 |
|
2.27 | 0.023 | 51.0% | 0.031 |
America | 1 |
|
<0.001 | 7.02 | NA | NA |
Subgroups of HF | ||||||
ICM patients | 5 |
|
1.46 | 0.145 | 88.9% | <0.001 |
IDCM patients | 8 |
|
4.55 | <0.001 | 0.0% | 0.504 |
Other HF patients | 14 |
|
5.28 | <0.001 | 78.6% | <0.001 |
Subgroup analysis of studies on zinc levels in HF patients versus control subjects stratified by geographic location.
Subgroup analysis of studies on zinc levels in HF patients versus control subjects stratified by HF subgroups.
The sensitivity analysis showed that no individual study had an extreme influence on the pooled effect (Table
The heterogeneity of the included studies through sensitivity analysis.
Excluded studies | SMD (95% CI) |
|
|
---|---|---|---|
Sullivan et al., 1979 |
|
72.90% | <0.001 |
Atlihan et al., 1990 (1) |
|
77.40% | <0.001 |
Atlihan et al., 1990 (2) |
|
76.30% | <0.001 |
Oster, 1993 |
|
77.20% | <0.001 |
Golik et al., 1993 (1) |
|
76.70% | <0.001 |
Golik et al., 1993 (2) |
|
77.30% | <0.001 |
Cénac et al., 1996 |
|
76.90% | <0.001 |
De Lorgeril et al., 2001 |
|
77.40% | <0.001 |
Topuzoglu et al., 2003 |
|
77.30% | <0.001 |
Kosar et al., 2006 (1) |
|
77.20% | <0.001 |
Kosar et al., 2006 (2) |
|
77.00% | <0.001 |
Kosar et al., 2006 (3) |
|
77.40% | <0.001 |
Kosar et al., 2006 (4) |
|
77.10% | <0.001 |
Kosar et al., 2006 (5) |
|
77.30% | <0.001 |
Kosar et al., 2006 (6) |
|
77.30% | <0.001 |
Salehifar et al., 2008 (1) |
|
77.10% | <0.001 |
Salehifar et al., 2008 (2) |
|
76.50% | <0.001 |
Salehifar et al., 2008 (3) |
|
77.30% | <0.001 |
Salehifar et al., 2008 (4) |
|
77.40% | <0.001 |
Salehifar et al., 2008 (5) |
|
77.20% | <0.001 |
Shokrzadeh et al., 2009 (1) |
|
76.30% | <0.001 |
Salehifar et al., 2009 (2) |
|
73.90% | <0.001 |
Salehifar et al., 2009 (3) |
|
77.20% | <0.001 |
Ghaemian et al., 2011 (1) |
|
73.30% | <0.001 |
Ghaemian et al., 2011 (2) |
|
67.60% | <0.001 |
Alexanian et al., 2014 (1) |
|
77.40% | <0.001 |
Alexanian et al., 2014 (2) |
|
76.80% | <0.001 |
Funnel plot of studies on zinc levels in HF patients versus control subjects.
Our meta-analysis included 12 reports with 1453 subjects from 27 case-control studies. The results showed that the serum zinc levels in HF patients were significantly lower than those in control subjects. This supported the supposition that there is some difference in zinc levels between HF patients and controls. In the subgroup analysis, the lower serum zinc levels in HF patients compared with control subjects were observed in different geographic locations (Europe, Asia, and America) and in IDCM and other HF patients. However, no significant difference in serum zinc levels was found in ICM patients, which could be due to the limited number of studies included in the analysis.
The mechanisms underlying the association between serum zinc levels and HF are still not fully understood. One underlying explanation for this meta-analysis showing is that micronutrient dyshomeostasis (such as zinc, copper, and Zn/Cu ratio dyshomeostasis) is associated with HF. The zinc and copper levels in the body affect each other, and decreased zinc levels are associated with deterioration of copper homeostasis and function [
Zinc is an important component of superoxide dismutase (SOD). Manganese SOD and copper-zinc SOD, as metal-binding proteins and enzymes, have highly efficient antioxidant mechanisms and a preventive effect on the occurrence of free radical induced injury. Zinc deficiency has been found to be associated with lower SOD activity [
According to Tousoulis et al., proinflammatory cytokines, such as interleukin-1 (IL-1), IL-6, and tumor necrosis factor-alpha (TNF-
In addition, HF may be associated with zinc deficiency through other mechanisms: reduced dietary intake (due to anorexia, nausea, and premature satiety with eating), reduced absorption due to gastrointestinal edema and impaired motility, increased intestinal zinc losses (protein-losing enteropathy), and excessive urinary excretion due to the use of diuretics [
To the best of our knowledge, this is the first meta-analysis to evaluate the association between zinc levels and HF. The sensitivity analysis showed that excluding any study from the pooled analysis did not vary the results substantially. Publication bias was also absent, as determined by funnel plot visualization and Begg’s test. However, the possible limitations of our study should be considered. First, some too-old studies were included in the meta-analysis, which might weaken the quality of the results. Second, heterogeneity could not be eliminated because of methodological diversity between studies; thus, the conclusion should be conservative. In addition, because of limited data, we did not analyze the zinc levels in other tissues; this might affect the comprehensive interpretation of the zinc levels in HF patients. Therefore, better designed studies are required to verify the results and further assess the role of zinc in the progress of HF.
In conclusion, the results of the meta-analysis indicate that there is a significant association between low serum zinc levels and HF. Meanwhile better designed studies are required to verify the results and further assess the role of zinc in the progress of HF.
The authors declare that they have no conflicts of interest.