Summary Electric arc welding is a technique widely employed across many industries, including metallurgy, construction, hydropower and the automotive industry. The main welding processes used in Quebec are gas metal arc welding (GMAW), flux cored arc welding (FCAW), metal cored arc welding (MCAW), shielded metal arc welding (SMAW) and gas tungsten arc welding (GTAW). Owing to the high temperatures reached at melting point, welding operations produce potentially toxic fumes consisting of particles and gases, the proportions and composition of which vary with several different factors, including the process used, the welding position and the type of pieces to be welded. The risks associated with these fumes and their constituents (primarily manganese [Mn], hexavalent chromium [Cr(VI)] and nickel [Ni]) include carcinogenicity, central nervous system damage and bronchopulmonary diseases. Welding fumes were classified as Group 1 carcinogens (carcinogenic to humans) by the International Agency for Research on Cancer (IARC) in 2017. Protecting workers from welding fumes is particularly important in Quebec, as some 24,000 Quebecers work in welding, and the standards and permissible exposure limits for welding fumes and constituents are often exceeded. At present, worker safety measures involve a combination of extraction of welding fumes at source, general ventilation and personal protective equipment. These approaches do not, however, adequately protect workers’ breathing zones in a number of work situations. Understanding the influence of welding parameters, such as type of electrode, its characteristics and shielding gas composition, can improve our ability to reduce worker exposure to welding fumes. The goal of this study was to contribute to the advancement of knowledge about the influence of welding parameters on fume concentrations and their metal constituents generated by electric arc welding operations. To this end, a document search in the form of a systematic literature review was conducted in bibliographic and factual databases and in the grey literature. Ten bibliographic databases and 12 online sources, including institutional and trade association websites, were queried for the period from 2000 to 2018. In all, 1,764 references were identified, 47 of which were selected. Forty-three of these references report on work carried out under experimental conditions and only four deal with work done in the field. In total, 21 welding parameters were assessed, leading to 85 assessment situations. (One assessment situation corresponds to the measurement of the influence of one welding parameter on the level of fumes generated, with respect to total fumes and specific contaminants.) The shielding gas and the characteristics of the electrode were the parameters most studied (each being associated with approximately 19% of the assessment situations), followed by voltage and welding processes (13% and 8% of assessment situations, respectively). Most assessment situations were related to the GMAW process (65%), followed by SMAW (18%), FCAW (11%) and lastly GTAW and MCAW (3% each). The results reported in the selected references show that the SMAW process generates the most fumes, while the GTAW process produces the least. For GMAW, using the pulsed and CMT™ (a short-circuit process that is a technical variant of GMAW) modes generates less Mn and Cr(VI) fumes than the conventional GMAW process. According to the results, however, although the SMAW process gives off more fumes, it generates fewer ultrafine particles (UFPs) than GMAW and GTAW do. Furthermore, the reported results clearly indicate that an increase in the carbon dioxide (CO2) fraction in a shielding gas mixture causes a corresponding increase in the quantity of fumes generated. An increase in voltage, current strength or electrode diameter also leads to higher fume levels, and the same thing occurs when more flux is used, in the electrode core (FCAW) or coating (SMAW). As for GMAW, pulsed spray transfer is associated with lower fume levels than the short-circuit and axial spray transfer modes. Overall, this study has shown that many welding parameters affect the generation of total fumes, specific contaminants and UFPs, by either reducing or increasing them. Field studies need to be conducted to arrive at an accurate assessment of the influence of what would appear to be the most important parameters. Special attention should be paid to the characterization of particle size, given the chemical nature of these particles and the potential risks to human health, especially to the respiratory and nervous systems.