Seasonality of Rotavirus Hospitalizations at Costa Rica’s National Children’s Hospital in 2010–2015

Abstract

Rotavirus is a leading cause of acute diarrhea in children worldwide. Costa Rica recently started universal rotavirus vaccinations for infants with a two-dose schedule in February 2019. We aimed to study the seasonality of rotavirus during the pre-vaccination era. We retrospectively studied a six-year period of hospital admissions due to rotavirus gastroenteritis. We estimated seasonal peak timing and relative intensities using trend-adjusted negative binomial regression models with the δ-method. We assessed the relationship between rotavirus cases and weather characteristics and estimated their effects for the current month, one-month prior and two months prior, by using Pearson correlation coefficients. A total of 798 cases were analyzed. Rotavirus cases predominated in the first five months of the year. On average, the peak of admissions occurred between late-February and early-March. During the seasonal peaks, the monthly count tended to increase 2.5–2.75 times above the seasonal nadir. We found the strongest negative association of monthly hospitalizations and joint percentiles of precipitation and minimal temperature at a lag of two months (R = −0.265, p = 0.027) and we detected correlations of −0.218, −0.223, and −0.226 (p < 0.05 for all three estimates) between monthly cases and the percentile of precipitation at lags 0, 1, and 2 months. In the warm tropical climate of Costa Rica, the increase in rotavirus hospitalizations coincided with dry and cold weather conditions with a two-month lag. The findings serve as the base for predictive modeling and estimation of the impact of a nation-wide vaccination campaign on pediatric rotaviral infection morbidity.

Rotavirus Seasonality: An Application of Singular Spectrum Analysis and Polyharmonic Modeling

“…For tropical climates, a meta-analysis of 26 studies conducted between 1975 and 2003 demonstrated a strong inverse relationship between monthly rotavirus incidence and climatological variables such as temperature, rainfall, and relative humidity [5]. In a recent study, we found significant correlation at a 2 month lag, indicating the delayed effects of cold and dry conditions on rotavirus incidence in Costa Rica [6]. Other studies have also reported negative lagged correlations between rotavirus and both temperature and precipitation in tropical and temperate regions of South Asia [3,13].…”

“…Rotavirus, one of the most common pediatric diarrheal diseases, typically peaks in cold and dry weather in both temperate and tropical climates [1,2,3,4,5,6,7,8,9,10,11,12]. However, the seasonal pattern and strength of the association between the incidence of rotaviral infection and ambient temperature, and the sensitivity of a seasonal peak of rotavirus to temperature vary.…”

“…However, the seasonal pattern and strength of the association between the incidence of rotaviral infection and ambient temperature, and the sensitivity of a seasonal peak of rotavirus to temperature vary. Some studies demonstrated a delay when the seasonal peak occurred 1–2 months after a nadir in ambient temperature [3,6,13]. It is not clear how stable the seasonal peak is and how strong the relationship between rotavirus infections and ambient temperature is in cold climates.…”

“…Some studies assessed the relationship with temperature by simply comparing cold and warm seasons [9,10,14]. In some studies, peaks of infections were presented as the month with highest counts [6], with occasional reporting of two seasonal peaks in RI [3,14]. Our understanding of what infection seasonality represents is currently limited by a lack of standardization of methodology for in-depth analysis of temporal variations in infections with marked seasonality and potential environmental factors that may act as proxies for seasonal change.…”

Effects of Data Aggregation on Time Series Analysis of Seasonal Infections

“…Seasonal infections, like respiratory and enteric infections, have well-known seasonal patterns which may vary by the pathogen [ 1 , 2 , 3 ], population [ 4 ], and location [ 2 , 5 ]. Common infections, such as giardiasis, have a single annual peak in incidence [ 6 , 7 ], while others, like cryptosporidiosis, demonstrate two peaks in certain locations [ 7 , 8 ].…”