Diurnal variations of fever in children: results from an app-based observational study

Fever is a frequent concern in early childhood. However, its management remains inconsistent, driven in part by persistent "fever phobia" and heterogeneous advice. Although body temperature follows a well-established circadian rhythm, the diurnal expression of fever in young children is insufficiently described. We aimed to characterise 24-hour patterns of febrile temperatures in children aged 0-7 years using real-world data from a participatory mobile health registry (FeverApp), with specific attention to age and antipyretic use.

We analysed 28,620 fever episodes recorded in the FeverApp registry between September 2019 and September 2023, comprising 12,454 children (median age 14 months) and 122,890 time-stamped temperature measurements. Mean hourly temperatures were calculated by pooling all eligible entries within hourly bins and were stratified descriptively by age group and recorded antipyretic use, without consideration of dependency. Between-group differences were summarised descriptively using mean differences.

Mean temperature demonstrated a clear diurnal pattern, reaching a nadir at 06:00 (38.25 °C) and a late-evening peak at 23:00 (38.93 °C), corresponding to an observed trough-to-peak difference of 0.68 °C. Infants (< 12 months) showed lower entry-level mean temperatures than older children (1-7 years), with differences smallest around the morning trough and more pronounced later in the day. Recordings close to documented antipyretic administration had higher mean temperatures throughout the day, consistent with confounding by indication, while the overall diurnal pattern remained preserved. High-fever measurements (≥ 39 °C) were least frequent in the early morning (05:00-08:59) and most frequent during late evening/night (21:00-00:59).

Fever in young children should be interpreted as a dynamic 24-hour pattern rather than as a single static temperature value. In this app-based observational cohort, recorded febrile temperatures followed a clear aggregate time-of-day pattern, with lower values in the early morning and higher values later in the day. This temporal context may help clinicians and caregivers interpret fever more realistically: a low morning value should not be used as reassurance in isolation, while late-day increases may partly reflect expected circadian dynamics. Because the analysis is descriptive and entry-level, these findings should not be interpreted as individual fever trajectories, child-level risks, or inferential estimates.

The FeverApp registry is registered in the German Clinical Trials Register (DRKS), number DRKS00016591, registered 07 February 2019 ( https://drks.de/search/en/trial/DRKS00016591 ), as a non-interventional observational study protocol.