rebound_events: Detect Rebound Hypoglycemia and Hyperglycemia

In cgmguru: Advanced Continuous Glucose Monitoring Analysis with High-Performance C++ Backend

Detect Rebound Hypoglycemia and Hyperglycemia

Description

Detects rebound hypoglycemia (Rhypo) and rebound hyperglycemia (Rhyper) as derived events built from cgmguru Level 1 event starts and a later opposite-threshold crossing within a configurable time window.

Rhypo is a Level 1 hyperglycemic event (>180 mg/dL for at least 15 minutes) followed by any glucose value <70 mg/dL within rebound_minutes. Rhyper is a Level 1 hypoglycemic event (<70 mg/dL for at least 15 minutes) followed by any glucose value >180 mg/dL within rebound_minutes. The later rebound side only needs a single qualifying threshold crossing.

These rebound hypoglycemia and hyperglycemia definitions are documented by Hansen and Bibby (2024); cgmguru applies them on its own event-preprocessed CGM grid.

Usage

rebound_events(df, type = c("all", "hypo", "hyper"),
 data_source = c("raw", "preprocessed"), reading_minutes = NULL,
 sort_time = FALSE, inter_gap = 45, rebound_minutes = 120,
 return_interpolated = TRUE)

Arguments

df	A dataframe containing continuous glucose monitoring (CGM) data with columns id, time, and gl.
type	Rebound direction to return. "hypo" returns Rhypo, "hyper" returns Rhyper, and "all" returns both.
data_source	Source interpretation for df. "raw" applies cgmguru event preprocessing first. "preprocessed" treats df as an already-preprocessed event grid and does not interpolate again.
reading_minutes	Time interval between readings in minutes. Can be a scalar, a vector matching nrow(df), or NULL. If omitted, it is inferred per id from positive timestamp differences.
sort_time	Logical. If TRUE, sort rows within each id by time. Defaults to FALSE.
inter_gap	Maximum gap in minutes to interpolate across when data_source = "raw". Defaults to 45.
rebound_minutes	Maximum bridge interval in minutes from the initial Level 1 event end to the later rebound threshold crossing. Defaults to 120.
return_interpolated	Logical. If TRUE, include the preprocessed event grid used for rebound detection as interpolated_data. Defaults to TRUE, so rebound_events() returns the preprocessed data by default.

Value

A list containing:

• events_total: Tibble with id, type, total_episodes, and avg_ep_per_day.

• events_detailed: Tibble with bridge boundaries (start_time, end_time, indices, and glucose values), the initial Level 1 event boundaries, rebound threshold crossing fields, and minutes_to_rebound.

• interpolated_data: The preprocessed event grid used for rebound detection, included by default when return_interpolated = TRUE, with columns id, time, and gl.

References

Hansen, K. W., and Bibby, B. M. (2024). Rebound hypoglycemia and hyperglycemia in type 1 diabetes. Journal of Diabetes Science and Technology, 18(6), 1392-1398.

See Also

detect_all_events, detect_hyperglycemic_events, detect_hypoglycemic_events, interpolate_cgm

Examples

df <- data.frame(
  id = "A",
  time = as.POSIXct("2026-01-01 00:00:00", tz = "UTC") + 0:8 * 5 * 60,
  gl = c(190, 195, 200, 170, 165, 160, 65, 100, 110)
)
rebound_events(df, type = "hypo", reading_minutes = 5)

cgmguru documentation built on July 8, 2026, 9:06 a.m.