Загрузка...

Tight glucose in ICU saved lives. Then it killed them. #Shorts

In 2001 Van den Berghe (NEJM) showed intensive insulin (80-110 mg/dL) reduced ICU mortality almost in half in a single-centre surgical ICU. International guidelines adopted the target. By 2009 NICE-SUGAR (NEJM, n=6,104) reported the opposite. Tight glucose target HR 1.14 (95% CI 1.02-1.28) for 90-day death vs conventional control, with a 14-fold increase in severe hypoglycemia. The target was retired.

Forest plot for ICU all-cause mortality:
- Van den Berghe 2001 (surgical ICU, n=1,548). RR 0.58 (95% CI 0.38, 0.87)
- Van den Berghe 2006 (medical ICU, n=1,200). RR 0.94 (95% CI 0.81, 1.09)
- NICE-SUGAR 2009 (mixed ICU, n=6,104). OR 1.14 (95% CI 1.02, 1.28)
- Pooled RE. RR around 1.05 (95% CI 0.91, 1.21). I-squared 78 percent

Meta-analysis lesson: WHEN I-SQUARED IS LARGE, THE POOLED ESTIMATE CAN BE MISLEADING. Three trials of the "same" intervention disagreed because the underlying populations were not the same (surgical, medical, mixed sepsis ICU). Pooling masks the U-shaped optimum. Too low causes hypoglycemia-driven deaths.

Practical takeaways:
1. Always report I-squared and tau-squared alongside the pooled estimate
2. If I-squared exceeds 50 percent, investigate before pooling. Subgroup by clinical features (population, dose, duration), not by drug class
3. When subgroups disagree in DIRECTION, pooling them is misleading
4. Use prediction interval, not just CI. PI tells you the likely effect in a NEW population

Sources:
- Van den Berghe G et al. NEJM 2001;345:1359-67 (Leuven surgical)
- Van den Berghe G et al. NEJM 2006;354:449-61 (Leuven medical)
- NICE-SUGAR Investigators. NEJM 2009;360:1283-97
- Griesdale DE et al. CMAJ 2009;180:821-7 (MA before NICE-SUGAR)
- Wiener RS et al. JAMA 2008;300:933-44

Tools: mahmood726-cyber.github.io/allmeta/
App walkthroughs: at 786-MIII Meta-analysis

#MetaAnalysis #EvidenceReversal #NICESUGAR #VanDenBerghe #ICUMedicine #Heterogeneity #Isquared #PoolingFallacy #SystematicReview #EBM

Видео Tight glucose in ICU saved lives. Then it killed them. #Shorts канала Evidence Reversal
Яндекс.Метрика
Все заметки Новая заметка Страницу в заметки
Страницу в закладки Мои закладки
На информационно-развлекательном портале SALDA.WS применяются cookie-файлы. Нажимая кнопку Принять, вы подтверждаете свое согласие на их использование.
О CookiesНапомнить позжеПринять