Open access Protocol Health effects of desert dust and sand storms: a systematic review and meta- analysis protocol Aurelio Tobias,1 Angeliki Karanasiou,1 Fulvio Amato,1 Marta Roqué,2,3 Xavier Querol1 To cite: Tobias A, Karanasiou A, Abstract Amato F, et al. Health effects Introduction Desert dust concentrations raise concerns Strengths and limitations of this study of desert dust and sand storms: about adverse effects on human health. During the last a systematic review and meta- ►► This systematic review will be the first one to re- decade, special attention has been given to mineral dust analysis protocol. BMJ Open trieve and evaluate published studies on the health particles from desert dust and sand storms. However, 2019;9:e029876. doi:10.1136/ effects of desert dust following a standardised pro- bmjopen-2019-029876 evidence from previous reviews reported inconclusive tocol for data collection and reporting of findings. results on their health effects and the biological ►► Prepublication history and ►► A meta-analysis will be used to provide evidence mechanism remains unclear. We aim to systematically additional material for this of the quantitative estimates of the health effects synthesise evidence on the health effects of desert dust paper are available online. To of desert dust exposures, accounting for relevant view these files, please visit and sand storms accounting for the relevant desert dust desert dust patterns from source areas, emissions, the journal online (http://​dx.​doi.​ patterns from source areas and emissions, transport and transport and composition. org/​10.​1136bmjopen-​2019-​ composition. ►► The diversity of methods to identify dust exposure 029876). Methods an analysis We will conduct a systematic events and the extensive health effects attributed to review that investigated the health effects of desert desert dust exposures might lead to study designs dust and sand storms in any population. The search Received 15 February 2019 with different methodological characteristics, which will be performed for any eligible studies from previous Revised 21 June 2019 will be carefully considered. reviews and selected electronic databases until 2018. Accepted 05 July 2019 ►► The results of this systematic review will provide ev- Study selection and reporting will follow the Preferred idence to fill the gap in the knowledge of the health Reporting Items for Systematic Reviews and Meta- effects of desert dust and may help to develop ap- Analyses guidelines. Data from individual studies will propriate preventive measures for dust episodes. be extracted using a standardised data extraction form. Quality of the studies will be assessed using a risk of bias tool for environmental exposures developed by experts convened by the WHO. A meta-analysis will be episodes is a complex issue. It can increase performed by calculating the appropriate effect measures particulate matter ambient concentrations by of association for binary and continuous outcomes from supplying large loads of mineral (or crustal) individual studies. Subgroup analyses will be performed by dust, but it can also carry anthropogenic geographical areas to account for desert dust patterns. pollutants, previously deposited in the source Ethics and dissemination No primary data will be areas or trapped by the high dust air mass collected. For this reason, no formal ethical approval during its atmospheric transport,5 6 and carry is required. This systematic review will help to fill the large amounts of microorganisms and toxic © Author(s) (or their research gaps in the knowledge of desert dust on human biogenic allergens.7 8 employer(s)) 2019. Re-use health. The results will be disseminated through a WHO Research on the health effects of desert permitted under CC BY. peer-reviewed publication and a conference presentation. dust has increased substantially over the last Published by BMJ. PROSPERO registration number CRD42018091809 1 Institute of Environmental decade. Although epidemiological studies Assessment and Water Research from affected areas suggest the potential (IDAEA), Spanish National Introduction health effects of desert dust, the evidence Research Council (CSIC), Background remains unclear, as well as the biological Barcelona, Spain Desert dust and sand storms play a significant mechanism. Previously published reviews, 2 Iberoamerican Cochrane Centre, Barcelona, Spain role in different aspects of weather, climate systematic or not, reported inconsistent 3 Biomedical Research Institute and atmospheric chemistry and represent results across studies and geographical Sant Pau (IIB Sant Pau), CIBER a severe hazard to the environment and regions.9–12 However, the reviewed studies Epidemiology and Public Health human health.1 2 Desert dust has a signifi- differed in terms of settings, dust exposure (CIBERESP), Barcelona, Spain cant impact on air quality, not only in areas assessment methods, outcomes and lagged Correspondence to close to the source points or regions but over exposures examined, and epidemiological Dr Aurelio Tobias; areas even a few thousands of kilometres study designs evaluated. For this reason, ​aurelio.​tobias@​idaea.​csic.​es distant.3 4 The air quality influence of dust none of the previous reviews could perform Tobias A, et al. BMJ Open 2019;9:e029876. doi:10.1136/bmjopen-2019-029876 1 Open access a meta-analysis to summarise the health effects of desert related to exposure to desert dust as mortality for all-nat- dust quantitatively. Thus, a systematic review with a proper ural causes and cause-specific (eg, cardiovascular and standardised protocol is required to provide a better respiratory) and hospital admissions, emergency room knowledge of health effects and mechanism of toxicity of admissions and emergency department specific causes. desert dust. Moreover, based on a previous evaluation of the human health disorders related to desert dust,14 we will also Objective consider cardiovascular and respiratory symptoms; lung The objective of our systematic review is to summarise diseases; coccidiomycosis; meningococcal meningitis; the evidence on the health effects of desert dust and sand dermatological disorder; and deaths or injuries resulting storms, properly accounting for the relevant desert dust from transport accidents. patterns from source areas and emissions, transport and composition. This systematic review will be carried out Types of studies in the framework of the WHO to inform the update of This systematic review will consider all types of relevant the WHO Global Air Quality Guidelines (AQGs), coor- studies published in peer-reviewed journals that assess the dinated by the WHO Regional Office’s European Centre health effects of desert dust and sand storms. However, for Environment and Health. the studies will be classified according to their observa- tional or experimental design (eg, quasi-experimental and randomised controlled trials). The observational Methods studies will also be classified at the ecological or indi- This protocol has been registered in PROSPERO and vidual level (eg, surveys, case–control, cohort studies). follows the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols checklist13 (online Information sources supplementary file 1). Any amendments to the protocol We will systematically conduct a comprehensive literature will be tracked and dated in PROSPERO. search. First, we will identify all studies already reported in previous reviews, systematic or not, on the health effects Patient and public involvement of desert dust and sand storms.9–12 Next, we will conduct a Patients and/or public will not be involved in this study, literature search for studies matching the study eligibility as this is a systematic review protocol. criteria in various databases, including MEDLINE (using Study eligibility criteria PubMed) and EMBASE. Finally, references of identified The WHO established the research question for this relevant studies will be scanned to identify additional systematic review at the third Meeting of the Global Plat- published data matching the study eligibility criteria. form on Air Quality and Health (Madrid, Spain, March The search strategy has been developed in collaboration 2017). with a health sciences librarian specialising in systematic search procedures. We used free-text for each search Population engine, considering different definitions and keywords This systematic review will include any human popula- for desert dust and sand storms and health outcomes tion, from developed and developing countries, living in (online supplementary file 2). We will retrieve eligible urban and in rural areas exposed to desert dust and sand published studies until 2018 without language restrictions storms. Vulnerable population subgroups to the effects and will translate the non-English abstracts and studies, if of desert dust and sand storms will be included, such as necessary. these with specific pre-existing health conditions (eg, respiratory or cardiovascular diseases), pregnant women, Selection of studies newborns, children or the elderly. Whenever applicable, Two reviewers will screen the titles and abstracts inde- the considered health effect of exposure to desert dust pendently, record and import search results of their find- in these vulnerable subgroups of the population will be ings in EndNote X7 (Clarivate Analytics). All studies will assessed separately. be classified for inclusion or exclusion based on the eligi- bility of this systematic review and further checked for Types of exposures duplication. Full-text articles will be retrieved if the study We will consider short-term and long-term exposure is unclear from the title and abstract, and the relevant to desert dust characterised by source apportionment, studies will be further confirmed for inclusion through component analysis, particle size and composition, use full-text review. The third reviewer will resolve discrepan- of weather data and back trajectories, or any other suit- cies between reviewers. able method. The exposure to desert dust should be compared with non-exposure to desert dust or exposure Data extraction to lower concentrations of desert dust. Reviewers will use a standardised data extraction form using Excel spreadsheet (Microsoft Corporation). Here, Types of outcome measures if a given study reports several health outcomes of interest, We will consider any measurable independent change in each outcome will take one record. Likewise, if a given the risk of adverse health effects (incidence or prevalence) study reports age or gender-specific effects, these will 2 Tobias A, et al. BMJ Open 2019;9:e029876. doi:10.1136/bmjopen-2019-029876 Open access also be collected. Moreover, if a given study reports more The measures of the association from individual studies than one effect estimate for each health outcome (eg, for will be summarised in a meta-analysis estimate, with its 95% several lags), the one that the authors favoured most will CI, based on fixed or random-effects model, depending be extracted. Main conclusion/findings for each study on the heterogeneity of the analysis.15 17 Statistical hetero- will also be collected as free text. The following character- geneity of the effect estimates between studies included istics of the included studies will be extracted: in the meta-analysis will be assessed using the I2 statistic, 1. Citation details (eg, first author, title, journal and date where values of 25%, 50% and 75% will be considered of publication). as a low, moderate and high degree of heterogeneity, 2. Study location (eg, city, country, region), study period, respectively.18 We will also report the between-study vari- characteristics of the study population (eg, age, gen- ance (τ2) and the Cochran Q test for heterogeneity.19 All der). analyses will be done using Stata Statistical Software V.15. 3. Study design and statistical analysis. Due to the relevance of the study setting and dust 4. Method to identify dust events and exposure levels for exposure assessment, subgroup analyses will be done by days with and without dust events. geographical areas, considering studies conducted in 5. Outcome assessment (eg, cause of mortality, morbidity Eastern Asia, Europe, Middle the East and other regions and other cause-specific outcomes). (eg, Australia and North America). Subgroup analyses 6. Confounders measured and adjusted for (eg, meteoro- by pre-existing health conditions, age groups (14, 15–64, logical variables and other air pollutants). ≥65 years) and gender (male and female) will also be 7. Effect measures of association, and its 95% CI, for days considered whenever data become available in the indi- with and without dust events. vidual studies. Quality assessment Assessment of reporting biases We will use a new domain-based risk of bias tool to assess Publication bias will be assessed when at least 10 studies the quality of the evidence across studies for associa- are included in the meta-analysis, using Begg’s rank tions between specific air pollutants and adverse health correlation test and examined using funnel plots.20 outcomes. The risk of bias tool, currently under develop- ment by a group of experts on environmental exposures Reporting conclusions convened by the WHO Guideline Development Group The systematic review will be conducted and reported (GDG) for the AQGs, will adapt the domains according in accordance to the Preferred Reporting Items for to the needs of the observational studies of exposure that Systematic Review and Meta-Analysis21 standards, with form the body of the evidence in air pollution. It will slight adaptations since these were originally intended be revised and adapted to by the systematic review team for healthcare intervention evaluation. We will base our members, and a member of the WHO GDG will support conclusions only on findings from the quantitative or the application of the tool. Two reviewers will assess the narrative synthesis of included studies for this systematic risk of bias independently, and disagreements will be review. The quality of the evidence will be assessed using resolved through discussion. an adaptation for environmental studies, convened by the GDG for the AQGs, of the Grading of Recommendations Data synthesis and statistical analysis Assessment (GRADE).22 Studies that undergo meta-anal- For binary outcomes (eg, mortality and morbidity), the yses will undergo GRADE assessment, but there might relative risk (RR) will be used as the standard effect also other studies included in the systematic review that measure of association across studies. The HR if reported, could not be used in the meta-analysis but will be used may be considered equivalent to RRs. If ORs are reported for developing conclusions. The conclusions from this in a given study and the outcome prevalence is >10% systematic review will be used to inform the update of the they will be approximated as RRs15 and similarly if the WHO AQGs, coordinated by the WHO Regional Office’s study design allows for it (eg, time-series and time-strati- European Centre for Environment and Health. fied case-crossover16). Meta-analysis input data will be as the percentage increase in the RR (eg, as %RR = (RR-1) Ethics and dissemination ×100%). When reporting effects of particulate matter for Formal ethical approval is not required, as no primary days with and without dust events the %RR will be stan- data will be collected. The results of this systematic review dardised for a given increment of 10 µg/m3, assuming a will fill the current research gaps in the knowledge on the linear exposure–outcome relationship. For continuous health effects of desert dust and sand storms and would outcomes, effect measures such as weighted mean differ- also help to alert the most sensitive population taking ence (WMD) or standardised mean difference (SMD) appropriate measures during dust episodes. All findings may be used. The scale of the available data will be will be shared and disseminated by WHO in a peer-re- primarily used to determine the choice of effect measure viewed publication and conference presentation. if studies report effect estimates with the same or similar scale (WMD) or when the outcome is measured using Acknowledgements Ivan Solà (Iberoamerican Cochrane Centre) for their guidance different scales (SMD).15 in reviewing the search strategy proposed in this protocol. Tobias A, et al. BMJ Open 2019;9:e029876. doi:10.1136/bmjopen-2019-029876 3 Open access Contributors AT is the guarantor. AT drafted the protocol and registered the subtropical Saharan air layer. Atmospheric Chemistry and Physics protocol in PROSPERO. MR and XQ reviewed and commented on the protocol in 2011;11:6663–85. PROSPERO. AK, FA, MR and XQ all reviewed and commented on this protocol. 7. H-M H, Rao CY, Hsu H-H, et al. Characteristics and determinants of ambient fungal spores in Hualien, Taiwan. Atmospheric Environment Funding The World Health Organization, supported by the Ministry of Foreign 2005;39:5839–50. Affairs of Norway, has funded this systematic review. 8. Griffin DW, Garrison VH, Herman JR, et al. African desert dust in the Caribbean atmosphere: microbiology and public health. Aerobiologia Competing interests None declared. 2001;17:203–13. Patient consent for publication Not required. 9. Hashizume M, Ueda K, Nishiwaki Y, et al. Health effects of Asian dust events: a review of the literature. Jpn. J. Hyg. 2010;65:413–21. Provenance and peer review Not commissioned; externally peer reviewed. 10. Karanasiou A, Moreno N, Moreno T, et al. Health effects from Sahara dust episodes in Europe: literature review and research gaps. Environ Data availability statement There are no data in this work. No data are available. Int 2012;47:107–14. Open access This is an open access article distributed in accordance with the 11. de Longueville F, Ozer P, Doumbia S, et al. Desert dust impacts on Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits human health: an alarming worldwide reality and a need for studies in others to copy, redistribute, remix, transform and build upon this work for any West Africa. Int J Biometeorol 2013;57:1–19. purpose, provided the original work is properly cited, a link to the licence is given, 12. Zhang X, Zhao L, Tong D, et al. A systematic review of global desert dust and associated human health effects. Atmosphere 2016;7:158. and indication of whether changes were made. See: https://​creativecommons.​org/​ 13. Shamseer L, Moher D, Clarke M, et al. Preferred reporting items for licenses/​by/​4.​0/. systematic review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explanation. BMJ 2015;349:g7647–g. 14. Goudie AS. Desert dust and human health disorders. Environ Int 2014;63:101–13. References 15. Higgins JPT, Green S, The Cochrane collaboration. Cochrane 1. Mahowald NM, Kloster S, Engelstaedter S, et al. Observed handbook for systematic reviews of interventions version 5.1.0 20th century desert dust variability: impact on climate and [updated March 2011], 2011: Accessed 15 Feb 2019. biogeochemistry. Atmos Chem Phys 2010;10:10875–93. 16. Lu Y, Zeger SL. On the equivalence of case-crossover and time 2. De Longueville F, Hountondji Y-C, Henry S, et al. What do we know series methods in environmental epidemiology. Biostatistics about effects of desert dust on air quality and human health in West 2007;8:337–44. Africa compared to other regions? Science of The Total Environment 17. Borenstein M, Hedges LV, Higgins JPT, et al. A basic introduction to 2010;409:1–8. fixed-effect and random-effects models for meta-analysis. Research 3. Prospero JM. Environmental characterization of global sources of Synthesis Methods 2010;1:97–111. atmospheric soil dust identified with the NIMBUS 7 total ozone 18. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta- mapping spectrometer (TOMS) absorbing aerosol product. Reviews analysis. Stat Med 2002;21:1539–58. of Geophysics 2002;40. 19. Cochran WG. The χ test of goodness of fit. Annals of Mathematical 4. Ginoux P, Prospero JM, Gill TE, et al. Global-scale Attribution Statistics 1952;23:315–45. of anthropogenic and natural dust sources and their emission 20. Begg CB, Mazumdar M. Operating characteristics of a RANK rates based on MODIS deep blue aerosol products. Reviews of correlation test for publication bias. Biometrics 1994;50:1088–101. Geophysics 2012;50:1–36. 21. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for 5. Mori I. Change in size distribution and chemical composition of kosa systematic reviews and meta-analyses: the PRISMA statement. BMJ (Asian dust) aerosol during long-range transport. Atmos Environ 2009;339(jul21 1:b2535–b. 2003;37:4253–63. 22. Guyatt GH, Oxman AD, Vist GE, et al. Grade: an emerging consensus 6. Rodríguez S, Alastuey A, Alonso-Pérez S, et al. Transport of on rating quality of evidence and strength of recommendations. BMJ desert dust mixed with North African industrial pollutants in the 2008;336:924–6. 4 Tobias A, et al. BMJ Open 2019;9:e029876. doi:10.1136/bmjopen-2019-029876