Inflammation is a multifactorial network of chemical signals in response to detrimental insults such as tissue injury, other noxious conditions and microbial infection. It is a critical immune response by the host that removes the harmful stimuli as well as the healing of the damaged tissue(Reference Ahmed1). Normal inflammation usually limits itself; however, dysregulation in any inflammatory factors can lead to abnormalities and pathogenesis. Systematic inflammation is believed to be associated with the initiation and progression of various chronic conditions such as cerebrovascular disease, dementia, ischemic heart and respiratory disease(Reference McLoughlin, Berthon and Jensen2). Also, the inflammatory response plays a vital role in different stages of tumour development, including initiation, progression, conversion to malignant and metastasis(Reference Grivennikov, Greten and Karin3).
The inflammatory processes involve activation of the immune system, directed migration of leucocytes, and release of pro-inflammatory cytokines and mediators(Reference Schwingshackl and Hoffmann4). Cytokines are molecules with glycoprotein or protein structure that affect interactions and communication between cells(Reference Zhang and An5). Among the various type of cytokines, TNF-α, IL-2, IL-6 and IL-1 are significant inducers of the acute-phase response(Reference Derosa, Maffioli and Simental-Mendia6,Reference Coussens and Werb7) . TNF-α through binding to its receptor (TNFR1 and TNFR2) regulates the cytokine cascade in inflammatory pathways as well as cell proliferation, survival, differentiation and apoptosis of immune cells(Reference Parameswaran and Patial8). IL-2 acts as both inflammatory and anti-inflammatory agent via binding to its receptor, and regulation of these dual effects is important in the treatment of many inflammatory diseases(Reference Banchereau, Pascual and O’Garra9). Membrane IL-6 receptor (mIL-6 R) mediates IL-6 actions including the differentiation and maturation of immune cells and the induction of acute-phase protein synthesis in hepatocytes(Reference Tanaka, Narazaki and Kishimoto10).
Production of these cytokines is mediated by different transcription factors (TF) including NF-κB, CCAAT/enhancer-binding protein (C/EBP)-β, the activator protein 1 (AP-1) and the nuclear factor of activated T cells(Reference Vaeth and Feske11–Reference Luo and Zheng14). Zn, as a trace element, plays an important role in the stabilisation of TF through Zn finger proteins(Reference Cassandri, Smirnov and Novelli15). A number of in vivo and in vitro studies have demonstrated that Zn can be effective in the regulation of mentioned TF, whereby it regulates inflammatory cascades(Reference Bao, Prasad and Beck16–Reference Kim, Aydemir and Cousins19).
Despite the number of trials which have investigated the effect of Zn supplementation on diseases associated with ongoing inflammation(Reference Jamilian, Foroozanfard and Bahmani20,Reference Khazdouz, Mazidi and Ehsaei21) as well as diabetes(Reference Cruz, de Oliveira and do Nascimento Marreiro22) and atherosclerosis(Reference Dias, Sena-Evangelista and de Oliveira Paiva23), there is no comprehensive systematic review and meta-analysis to obtain a conclusive finding on the impact of Zn supplementation on cytokines. Therefore, this systematic review and meta-analysis study is conducted to evaluate the potential anti-inflammatory effects of Zn supplementation.
Material and methods
Search strategy
This systematic review and meta-analysis study was performed and reported in accordance with the guiding principle and recommendation of the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA)(Reference Moher, Liberati and Tetzlaff24). PubMed (Medline), Scopus, Web of Science and Embase databases were searched up to 10 December 2020 by three authors (M. Z., Y. K. H. and A. H. F.) independently with no limitation for year and language. Search was performed by using following search pattern: (‘Zinc’ [Mesh] OR zinc[Title/Abstract]) AND (Interleukin-8[Title/Abstract] OR IL-8[Title/Abstract] OR CXCL8[Title/Abstract] OR NAP[Title/Abstract] OR NAP1[Title/Abstract] OR IL-2[Title/Abstract] OR interleukin-2[Title/Abstract] OR Interleukin-1[Title/Abstract] OR IL-1[Title/Abstract] OR Interleukin-1beta[Title/Abstract] OR Interleukin-1β[Title/Abstract] OR IL-1 β [Title/Abstract] OR IL-1beta[Title/Abstract] OR ‘Interleukin-2’ [Mesh] OR ‘Interleukin-1beta’ [Mesh] OR ‘Interleukin-8’ [Mesh] OR ‘Tumor Necrosis Factor-alpha’ [Mesh] OR tumor necrosis factor-alpha[Title/Abstract] OR TNF[Title/Abstract] OR ‘tumor necrosis factor’ [Title/Abstract] OR ‘tumor necrosis factor-α’ [Title/Abstract] OR tnf-alpha[Title/Abstract] OR tnf- α[Title/Abstract] OR tnfα[Title/Abstract] OR Tumour necrosis factor[Title/Abstract] OR ‘Interleukin-6’ [Mesh] OR interleukin-6[Title/Abstract] OR IL-6[Title/Abstract] OR interleukin6[Title/Abstract] OR IL6[Title/Abstract] OR inflammation[Title/Abstract]) AND (randomized controlled trial[Publication Type] OR controlled clinical trial[Publication Type] OR ‘clinical trial’ [Title/Abstract] OR random * [Title/Abstract] OR supplementation[Title/Abstract] OR placebo[Title/Abstract] OR groups[Title/Abstract] OR trial[Title/Abstract] OR ‘randomized controlled trial’ [Title/Abstract] OR ‘controlled clinical trial’ [Title/Abstract]. We used the wild-card term ‘*’ to increase the search sensitivity.
Study selection and inclusion and exclusion criteria
After removing duplicate records, the titles and abstracts of the searched studies were screened based on the inclusion and exclusion criteria by three authors (O. M. T. R., E. F. and M. A.) independently. Controlled trials with parallel or cross-over design which have investigated the effects of Zn supplementation on the inflammatory cytokines were included. The PICO strategy for the research question of the study was Patient/Population (P): subjects with age of ≥15 years; Intervention (I): oral supplementation with Zn; Comparison (C): placebo or control group; and Outcome (O): changed TNF-α, IL-2 or IL-6 serum/plasma levels. Exclusion criteria were considered as follows: (i) other types of studies (in vitro, in vivo, ex vivo, quasi-experimental, reviews, letters, conference abstracts, case reports and observational studies), (ii) Zn supplementation along with another ingredient, (iii) infants and juvenile subjects, and (iv) lack of adequate information. Also, the reference lists of included studies were evaluated for other potential articles.
Data extraction
Articles meeting the inclusion criteria were abstracted by two authors (A. H. F. and L. P.) independently on the following items: first author and year of publication, journal, the region of investigation, sample size in each group, sex of subjects, mean age of subjects in each group, administered dosage and form of Zn, duration of treatment, and serum/plasma levels of inflammatory markers before and after the trial in each group. Any disagreements were discussed and resolved with a third reviewer (B. A.).
Study quality assessment
The Cochrane Collaboration tool was used for quality assessment of each trial included by two authors(Reference Higgins, Altman and Gøtzsche25). This tool assesses random sequence generation, allocation concealment details, blinding, elucidating of dropouts, reporting bias and other possible causes of bias. Each parameter was reported as high (−) or low (+) risk of bias, and unclear (?).
Statistical analysis
All data in included studies were reported as mean ± standard deviation. Reported means and standard errors, medians and ranges, and medians and interquartile ranges (Q25–Q75) were converted to means and standard deviations with statistical calculations. Pooled effect sizes were calculated using a random effect model with restricted maximum likelihood method. Because of the homogenised measurement units, the pooled effect sizes were expressed as standardised mean differences (SMD) with 95 % CI. I 2 statistic was used for the assessment of between-study heterogeneity and I 2 > 50 % was reported as high heterogeneity(Reference Higgins, Thompson and Deeks26). Potential sources of heterogeneity were identified using meta-regression and subgroup analysis based on sex, administered Zn dosage and form, duration, study population, and mean age. In meta-regression analysis, any linear relationship between the effect sizes and intervention duration, dose, and sample size were assessed. The influence of single study removal on the pooled effect size was assessed using sensitivity analysis. Publication bias was assessed using visual inspection of the funnel plot. In the presence of funnel plot’s asymmetry, the trim and fill method was carried out for adjusting the results with estimating the missing studies that might exist in pooled analysis and the effect of these studies on the outcome. Small-study effects were investigated using Begg’s adjusted rank correlation and Egger’s regression asymmetry tests(Reference Begg and Mazumdar27,Reference Egger, Davey Smith and Schneider28) . In all tests, P < 0·05 was considered as significant level. All statistical analyses were performed using STATA version 16.0 (Stata Corporation).
Results
Study selection
Totally, 10 676 articles were obtained from a search in databases. Titles and abstracts of 7545 studies were screened after eliminating duplicate articles. Among which, 178 articles were eligible for assessing full texts. Finally, a total of twelve articles met the inclusion criteria and included in the quantitative synthesis. PRISMA flow chart for literature search and selection is presented in Fig. 1.
The process of study selection through stages shown by Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) flow chart. RCT, randomised controlled trial.
Characteristics of selected studies
Among included studies, eight studies with nine treatment arms, eight studies with eight treatment arms and two studies with three treatment arms investigated the effect of Zn supplementation on TNF-α, IL-6 and IL-2 levels, respectively. Four of the included studies were performed in Iran(Reference Ranjbar, Shams and Sabetkasaei29–Reference Roshanravan, Tarighat-Esfanjani and Alamdari32). Other studies have been carried out in Turkey(Reference Kara, Ozal and Gunay33), Brazil(Reference de Moura, Soares and de Lima Barros34), Russia(Reference Freiberg, Cheng and Gnatienko35), Australia(Reference Foster, Petocz and Samman36), Thailand(Reference Meksawan, Sermsri and Chanvorachote37), USA(Reference Bao, Prasad and Beck38) South Korea(Reference Kim and Ahn39) and Poland(Reference Suliburska, Skrypnik and Szulińska40) with a total sample size of 680 (varied from 17 to 254 participants). The included studies were performed from 2010 to 2020. The mean age of subjects ranged between 15 and 67 years. The Zn salts gluconate and sulphate were used in the included studies which provided 12–50 mg/d of elemental Zn. The duration of supplementation was between 4 and 72 weeks. Participants of four studies were chosen from females and others from both sexes. Inflammatory cytokines in five and seven studies were measured in serum and plasma, respectively. Characteristics of included studies are outlined in Table 1.
Characteristics and baseline measurements of included studies
(Mean values and standard deviations)
INT, intervention group; CONT, control group; NR, not reported; Sulph, sulphate; Gluc, gluconate; F, female; M, male; PW, postmenopausal women; PCOS, polycystic ovary syndrome; IGT, impaired glucose tolerance.
* From the same studies with investigating Zn supplementation on different subjects.
† Expressed as range.
Risk of bias assessment
The results of Cochrane Collaboration’s tool on the assessment of the risk of bias are presented in Fig. 2. As shown, random allocation was not observed in one included study. Moreover, the intention to treat protocol for analysing data in trials was not performed in four included studies. Four included studies had higher quality compared with others.
The result of risk of bias assessment using Cochrane Collaboration’s risk of bias tool: each risk of bias item for included studies (green (+) means low risk of bias, yellow (?) means unclear risk of bias, red (−) means high risk of bias).
Effects of zinc on TNF-α
Zn supplementation had not a significant effect on TNF-α level in pooled estimate (SMD = 0·42 pg/ml; 95 % CI −0·31, 1·16; P = 0·257) (Fig. 3(a)). This result was confirmed by sensitivity analysis. However, the gluconate form of Zn supplement led to a significant decrease in TNF-α level following subgroup analysis (SMD = −0·89 pg/ml; 95 % CI −1·35, −0·43; P < 0·001) (Table 2). Moreover, TNF-α reduction was shown following <40 mg/d elemental Zn supplementation (SMD = −0·37 pg/ml; 95 % CI −0·74, −0·00; P = 0·048). However, ≥40 mg/d elemental Zn supplementation resulted in a significant increase in TNF-α serum/plasma level (SMD = 2·73 pg/ml; 95 % CI 0·08, 5·39; P = 0·044). There was a significant between-study heterogeneity (I 2 = 87·9 %, P < 0·001), which was reduced with subgrouping by the mean age of participants, Zn salt, study population and duration (Table 2). Meta-regression analysis showed that the effect size had no significant relationship with duration, mean age and sample size. Begg’s test did not confirm the presence of a small-study effect (P = 0·118). However, a significant small-study effect was shown following Egger’s test (P = 0·02). Moreover, visual inspection showed an asymmetric distribution in the funnel plot (Fig. 3(b)). However, trim and fill analysis confirmed the observed result (SMD = 0·98 pg/ml; 95 % CI −1·37, 3·34; P > 0·05).
Forest plot (a) detailing standardised mean differences (SMD) and 95 % confidence intervals and funnel plot (b) displaying publication bias in the studies reporting the effects of zinc supplementation on TNF-α level.
Pooled estimates of effects of zinc on inflammatory markers within different subgroups
(Standard mean differences (SMD) and 95 % confidence intervals)
F, female; M, male; NR, not recognised.
Effects of zinc on IL-2
Pooled analysis revealed that Zn supplementation had no significant effect on IL-2 level (SMD = 1·64 pg/ml; 95 % CI −1·31, 4·59; P = 0·277) (Fig. 4(a)). However, sensitivity analysis revealed that removing Beserra de Moura et al. study(Reference de Moura, Soares and de Lima Barros34) in pooled analysis led to a significant increase in IL-2 level following Zn supplementation (SMD = 2·96 pg/ml; 95 % CI 2·03, 3·88; P < 0·05). Because of limited number of studies on IL-2, subgroup analysis and meta-regression were not possible. Egger’s test but not Begg’s test showed publication bias due to small-study effects (P = 0·024 and P = 0·296, respectively). Moreover, visual inspection of the funnel plot illustrated the asymmetric distribution of studies (Fig. 4(b)). However, trim and fill analysis with three observed studies confirmed the observed results (SMD = 1·55 pg/ml; 95 % CI −0·97, 4·08; P > 0·05).
Forest plot (a) detailing standardised mean differences (SMD) and 95 % confidence intervals and funnel plot (b) displaying publication bias in the studies reporting the effects of zinc supplementation on IL-2 level.
Effects of zinc on IL-6
The results of forest plot showed that Zn supplementation could decrease IL-6 level (SMD = −0·76 pg/ml; 95 % CI −1·28, −0·24; P = 0·004) (Fig. 5(a)). Duration of supplementation, study population and administered dosage were recognised as sources of observed high between-study heterogeneity (I 2 = 85·1 %, P < 0·001) (Table 2). Also, subgroup analysis revealed that a higher dosage of Zn supplements (≥40 mg/d), female sex, healthy population and higher mean age (>40 years) resulted in a more reduction of IL-6 levels (Table 2). Among Zn forms, zinc gluconate had an ameliorative effect on the IL-6 level (Table 2). Meta-regression analysis revealed no significant relationship between the effect size and intervention duration, dose, and sample size. There were no small-study effects following Egger’s and Begg’s tests (P = 0·181 and 0·108, respectively). Moreover, symmetric distribution of studies was visualised in the funnel plot (Fig. 5(b)).
Forest plot (a) detailing standardised mean differences (SMD) and 95 % confidence intervals and funnel plot (b) displaying publication bias in the studies reporting the effects of zinc supplementation on IL-6 level.
Discussion
As our knowledge, this systematic review and meta-analysis study is the first comprehensive investigation evaluated the possible effects of Zn supplementation on serum/plasma profile of inflammatory cytokines including IL-2, IL-6 and TNF-α using published trial data. We revealed that Zn supplementation had an ameliorative effect on the IL-6 level. Zn forms, dosage, intervention duration, study population and mean age of participants could affect the pooled estimate and considered as the potential sources of heterogeneity. We did not include C-reactive protein (CRP) level in our meta-analysis since other meta-analysis study by Mousavi et al. revealed that Zn supplementation could decrease CRP level(Reference Mousavi, Djafarian and Mojtahed41).
We demonstrated that Zn supplementation had no ameliorative effect on TNF-α level following pooled analysis of nine relevant studies. Subgroup analysis revealed that intervention duration, study population, mean age of participants and Zn forms of the supplement were potential sources of high heterogeneity. Duration of supplementation and the mean age of participants were from a broad range, and these could cause various results. Analysis of two studies that administered the gluconate form revealed that zinc gluconate had a beneficial effect on the TNF-α level. One possible cause is difference in bioavailability. Sapota et al. revealed that the rats administered zinc sulphate had the lowest Zn in the prostate tissue compared with those supplemented with zinc gluconate(Reference Sapota, Daragó and Skrzypińska-Gawrysiak42). On the other hand, another in vivo study reported that zinc gluconate and zinc sulphate had equivalent bioavailability(Reference Zhang, Yu and Zhang43). The second cause is study population; in one study administered zinc gluconate, subjects had Zn deficiency(Reference de Moura, Soares and de Lima Barros34). It seems Zn supplementation had a more beneficial effect on inflammatory markers in the Zn deficiency status. In the other study, subjects were obese persons who had altered Zn redistribution(Reference Kim and Ahn39). Surprisingly, Zn supplementation increased TNF-α level in higher dosages (≥40 mg/d elemental Zn) subgroup and decreased TNF-α level in lower dosage (<40 mg/d elemental Zn). The tolerable upper intake level of Zn for adults has been determined 40 mg/d(44). Zn acts as a pro-oxidant agent in overload condition through disrupting of mitochondria homoeostasis and excessive reactive oxygen species production(Reference Lee45). Moreover, Zn toxicity might be related to increase in pro-inflammatory cytokine production(Reference Plum, Rink and Haase46).
We failed to find a significant association between Zn supplementation and the IL-2 production in the pooled estimate. However, sensitivity analysis revealed that Zn supplementation led to a significant increase in IL-2 level after omitting Beserra de Moura et al. study(Reference de Moura, Soares and de Lima Barros34). The intervention groups in the mentioned study were subjects with Zn deficiency in whom their serum Zn levels were not raised enough to reach Zn sufficiency status during the supplementation, likely because of short duration and low dosage of intervention. Since IL-2 production is impaired in Zn-deficient subjects(Reference Prasad, Beck and Kaplan47), their IL-2 levels remained low. IL-2 reduction in Zn deficiency status is related to the expression of the TF cAMP-responsive element modulator α (CREMα) which is involved in IL-2 transcription. CREMα binding site is cAMP-responsive element which is located within the IL-2 promoter. Therefore, its high expression inhibits IL-2 transcription(Reference Kloubert, Wessels and Wolf48). In vitro investigation revealed that Zn deficiency increased CREMα expression(Reference Kloubert, Wessels and Wolf48).
IL-2 has dual effects on inflammatory pathways. It is a promoter of T-cell proliferation and T-helper 1 (Th1) and Th2 effector cells generator and whereby induces inflammation. On the other hand, IL-2 inhibits the production of inflammatory Th17 cells and has a pivotal role in regulatory T-cell maintenance(Reference Hoyer, Dooms and Barron49). In lower dosage, IL-2 has been used as a therapeutic agent in the various types of cancer and inflammatory conditions because of its effect on immune system function and anti-inflammatory impacts in lower dosage(Reference Banchereau, Pascual and O’Garra9,Reference Shachar and Karin50) . However, in higher dosage, IL-2 may lead to toxicity and inflammation in some organs like skin and lung(Reference Shachar and Karin50). It is not known whether observed increase in IL-2 concentration in our meta-analysis has inflammatory effects or anti-inflammatory. It must be noted that participants of included studies in our meta-analysis on the effect of Zn treatment on IL-2 level were from people with a low concentration of serum/plasma Zn. Moreover, other ex vivo studies revealed that Zn supplementation had increasing effect on IL-2 production in Zn-deficient subjects(Reference Prasad, Beck and Kaplan47,Reference Rahfiludin, Wirjatmadi and Agusni51) . Therefore, it seems that Zn supplementation corrected the low level of IL-2 as an anti-inflammatory agent.
Nine comparisons of eight relevant studies revealed that Zn supplementation could decrease the IL-6 level. Intervention duration, dosage, study population and Zn forms were potential sources of high heterogeneity. Similar to TNF-α pooled analysis, intervention duration and elemental Zn dosage were from a broad range: 12–50 mg/d and 8–72 weeks, thereby they could cause heterogeneity in the results. Higher dosage of elemental Zn (≥40 mg/d), gluconate form and older subgroups had more decrease of IL-6 serum/plasma level. As discussed above, more beneficial effect of gluconate form may be related to its better bioavailability. Elderly people are more at risk for chronic inflammation and oxidative stress than younger people(Reference Prasad, Bao and Beck52). As a result, Zn supplementation seems to have a more beneficial effect on inflammation in the elderly subjects than in the young. Higher dosage of Zn had different impacts on TNF-α and IL-6 levels. TNF-α is an early response cytokine(Reference Mizgerd, Spieker and Doerschuk53,Reference Andreasen, Krabbe and Krogh-Madsen54) . Pro-oxidant conditions, caused by Zn overload, firstly elevate TNF-α level. It seems that a long-term supplementation with a high dosage of Zn is needed to have an increasing effect on IL-6 levels. The results on healthy and patient subjects were not very different, and more studies on people with different health statuses are needed to draw the right conclusion. Bao et al. study(Reference Bao, Prasad and Beck38) was the only study with higher baseline plasma Zn. However, sensitivity analysis revealed that omission of mentioned study did not change the overall result.
Zn has a possible effect on IL-2, IL-6 and TNF-α levels through the regulation of NF-ĸB activation. Dimerisation of NF-ĸB mediated by other subunits like p50, p65, c-Rel and RelB is essential for its binding to DNA(Reference Giuliani, Bucci and Napolitano55). Zn finger proteins like A20, growth factor independence-1 (Gfi-1), Zn finger and BTB domain containing 20 (ZBTB20), and Zn finger protein-64 (ZFP-64) are involved in NF-ĸB signalling; as A20 and Gfi-1 inhibit NF-ĸB dimerisation; ZBTB20 and ZFP-64 promote NF-ĸB dimerisation(Reference Faghfouri, Zarrin and Maleki56). It has been found that Zn supplementation induces A20 binding to mRNA and DNA via the upregulation of mRNA and DNA-specific sites and whereby acts its anti-inflammatory effect(Reference Prasad, Bao and Beck52). In an in vivo study, Zn was contributed to the reduction of NF-κB p65 mRNA expression in the jejunum of weaned piglets(Reference Hu, Cheng and Li57). Moreover, stabilisation of other TF including CCAAT/enhancer-binding protein (C/EBP)-β, AP-1 and nuclear factor of activated T cells which are contributed in the production of cytokines is mediated by Zn. Zn status also can be effective in mRNA expression of a cytokine through binding to metal response elements on the promoter of target genes(Reference Bao, Prasad and Beck58,Reference Cousins59) .
Some limitations of our study must be noted. First, the quality of some included studies was low. Random allocation and blinding of participants, researchers, and outcome assessments were observed only in six and four included studies, respectively. Second, present review was not registered in any registration databases. Third, since some studies had not determined the basic serum/plasma level of Zn in study population, subgroup analysis based on basic Zn level was not performed. Fourth, as study population of included studies were different, we used random effect model to control unobserved heterogeneity. However, random effect model does not generalise the results of the performed meta-analysis to real-world situations(Reference Ades, Lu and Higgins60). Fifth, because of limited number of studies on IL-2, subgroup analysis and meta-regression were not possible. Therefore, sources of heterogeneity were unknown and this question the generalisability of our result on IL-2. Sixth, we could not include studies on other inflammatory markers. There are a variety of inflammatory cytokines like IL-1 and IL-8 which we included in our search regimen. However, because of the limited number of studies on each, meta-analysis on them was impossible. Other trial studies on the other inflammatory markers are needed to find a comprehensive conclusion about the effect of Zn on the inflammatory process.
There are some strengths in our meta-analysis. First, it is the first comprehensive meta-analysis study which evaluated the effect of Zn supplementation on the inflammatory cytokines. Second, we analysed the included studies based on different subgroups to find a defined conclusion. Third, we assessed all included studies with appropriate tests to find any source of heterogeneity and bias between the studies.
Conclusion
Zn supplementation can decrease IL-6 serum/plasma level. Zn has no significant effects on IL-2 and TNF-α production. Lower dosage of zinc gluconate has ameliorative effect on TNF-α serum/plasma level. Dosage and form of Zn supplement are two key factors in the effectiveness of Zn supplementation in modifying inflammatory responses. A dosage lower than upper intake level of zinc gluconate seems to have a better effect on inflammatory markers.
Acknowledgements
The research protocol was approved by Vice Chancellor for Research (VCR), Tabriz University of Medical Sciences (registration code: 64693).
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Conceptualisation: A. H. F. and B. A. Database searching: M. Z., A. H. F. and Y. K. B. Screening: O. M. T.-R., E. F. and M. A. Data extraction: A. H. F. and L. P. Drafting of the paper: A. H. F. and Y. K. B. Statistical analysis: M. Z. Critical revision: B. B., A. K. and B. A. All the authors approved the final version to be submitted.
The authors declare that there are no conflicts of interest.
