Moderators
Our first analysis aimed to determine whether pathogen exposure in selected studies led to reduced host fitness without environmental stressors. For this analysis, we used the response trait category (fecundity or survivorship) as the only moderator and included only data from hosts in control (i.e., no environmental stress) conditions. We then focused on effects of environmental stressors on host-pathogen dynamics and examined three factors that could moderate the magnitude of these effects. For Q1, we considered infection status (infected and uninfected), stressor type, and response trait (fecundity and survivorship) as moderators. We were specifically interested in whether infection status amplified any negative fitness consequences of stressors. As mentioned above, stressors were of three types: 1) endogenous environmental factors (e.g., temperature, humidity, salinity, dissolved oxygen, and habitat structural complexity), 2) chemical pollution by toxins or synthetic compounds typically derived from pesticides or herbicides, and 3) resource limitation (restricted access to food or specific nutrients, like nitrogen and phosphorus). For response traits, fecundity was typically recorded as total number of offspring, whereas survivorship was reported as proportion alive, number alive, and sometimes, time to death.
In Q2, we focused exclusively on infected individuals under the abovementioned criteria. We investigated stressor type and response trait as moderators. We aimed to contrast effects of stress on fitness vs. infectivity responses. We, therefore, included two additional response traits as infectivity proxies: infection intensity and prevalence. Prevalence was always reported as number or percentage of infected individuals. Infection intensity was often quantified in different ways for different types of pathogens, for example, (log) copy number for viruses, colony-forming units for bacteria, mean number of cercaria for helminths, and spore counts for fungi. To compare relative sensitivity of fitness and infectivity, and because prevalence and infection intensity represent the opposite of host defense, signs of unbiased standardized mean differences were flipped. By doing so, a positive effect size reflects greater defense and a beneficial outcome for hosts, whereas, for fitness traits, a positive sign indicates higher survivorship or fecundity.
We complemented these main models for Q2 with two additional moderators in separate analyses. We investigated whether transmission environment (terrestrial or aquatic) or transmission mode (direct or indirect) modulates effects of environmental stressors on infectivity and host fitness responses. For hosts that occupy different environments across life stages, we categorized transmission environment based on life stage of hosts exposed or infected in each study, which was typically the most susceptible life stage to the target pathogen. We classified pathogen transmission as ‘indirect’ if it met one of three conditions: 1) pathogen required an ecologically distinct intermediate host to complete its life cycle, 2) pathogen was transmitted between ecologically similar hosts via vectors, or 3) pathogen could survive independently of host during free-living stage. Otherwise, pathogens were considered as having ‘direct’ transmission between ecologically similar hosts.