Regional and type-specific variations in the global seasonality of human parainfluenza viruses and the influence of climatic factors: a systematic review and meta-analysis

Abstract

Background: Human parainfluenza viruses (hPIVs) are common viral causes of acute respiratory infections, resulting in substantial global disease burden. Seasonal patterns of hPIV epidemics can vary by geographical region and viral type, although these patterns are not well understood at a global level. We aimed to characterise regional and type-specific variations in hPIV seasonality and assess the potential role of climatic factors in explaining these variations.

Methods: In this systematic review and meta-analysis, we collected monthly aggregated seasonal activity data for hPIV and its four types (hPIV-1, hPIV-2, hPIV-3, and hPIV-4) from various sources, including a systematic search of Embase, MEDLINE, and Ovid Global Health, for literature published between Jan 1, 2000, and Dec 31, 2024; unpublished data contributed by an established collaborative network; and public viral surveillance datasets from online platforms. We included studies that continuously tested hPIVs throughout their study period and reported seasonal activity in defined geographical locations on a monthly basis (or if monthly data could be derived from reports). We excluded published studies if they had fewer than 20 cases, focused on specific medical conditions, or contained duplicate data from published literature or publicly available datasets. A prespecified collection template was used to collect data from members in the collaborative network. We extracted site-specific monthly case counts of combined hPIVs and each viral type and used the annual average percentage approach to assess relative circulating strength, epidemic onset and peak month, and epidemic duration by virus type and latitude. We identified type-specific transmission zones of countries with similar circulating patterns with the k-means method. A local regression model (selected by leave-one-out cross-validation) was used to explore climatic factors associated with variations in hPIV monthly circulating activity. The study was registered with PROSPERO, CRD42023370261.

Findings: We included 115 records in total: 103 studies identified from the published literature, five studies contributed by collaborators, and data from seven public surveillance datasets. We included 306 719 cases from 141 sites in 64 countries. We found that hPIV-3 exhibited distinctive seasonal patterns compared with the other three hPIV types. In temperate regions, hPIV-3 seasons typically occurred in spring, summer, and winter, with a median onset in April (IQR April-May) in the northern temperate region and July (July-July) in the southern temperate region. hPIV-1, hPIV-2, and hPIV-4 seasons typically occurred in autumn, winter, and summer, with median onsets between August and October in the northern temperate region and between April and May in the southern temperate region. Both epidemic onset and peak timing for hPIV-1, hPIV-2, and hPIV-4 were less consistent in tropical and subtropical regions than in temperate regions, whereas the seasonality of hPIV-3 remained generally consistent across regions. Northern temperate and subtropical countries typically clustered in shared transmission zones for hPIV-1, hPIV-2, and hPIV-3 with a few exceptions, as did countries in the southern hemisphere. Nevertheless, hPIV-1 and hPIV-4 peak timings were delayed as latitude increased in the northern hemisphere (Pearson's r=0·62 [p=0·0012] for hPIV-1 and r=0·53 [p=0·049] for hPIV-4). Type-specific climate models yielded better fits (with greater area under the receiver operating characteristic curve values) than models for combined hPIVs. In temperate regions, higher hPIV-1, hPIV-2, and hPIV-4 activity correlated with declining temperature and increasing relative humidity (all p values <0·0001), whereas higher hPIV-3 activity was correlated with rising temperature (r_s=0·61; p<0·0001). In subtropical and tropical regions, the climate models showed suboptimal performance. Exploratory analyses showed differential timing shifts in hPIV epidemics across six included countries following the lifting of COVID-19 non-pharmacological interventions.

Interpretation: Our results characterise both between-type and regional variations in hPIV seasonality and the differential effects of monthly temperature variability and relative humidity on the global seasonality of different hPIV types. These findings have important implications for development of global hPIV surveillance and epidemic prediction in diverse locations. Substantial gaps in hPIV type-specific seasonality data remain in many countries, highlighting the need to expand surveillance to improve characterisation and prediction of hPIV epidemics.

Funding: National Natural Science Foundation of China.