Thread by @leahkosyakovsky

Our #systematicreview and #metaanalysis on the association of #sepsis with long-term #CVD is out now in @yourICM ! Surviving sepsis was associated with ⬆️#cvMI, #heartfailure, and #stroke over 5+ years. Take-home points below. tinyurl.com/p5n634jy

1. Many studies have examined the risk of #CVD after #sepsis, but variable population and outcome selection has complicated generalizability. We estimated the risk of #cvMI, #heartfailure, #stroke, and composite #CVD after sepsis relative to hospital and population controls.

2. 12,649 citations identified, 27 (25 cohort, 2 case-crossover) studies included. Median of 4,289 (IQR 502–68,125) sepsis survivors and 18,399 (IQR 4,028–83,506) controls per study. We assessed cumulative incidence and relative risk of #CVD from 30 days to longest follow-up.

3. The distribution of maximum follow-up in the included studies was as follows: 30 days to less than 1 year (7 studies), 2 years (2 study), 3 years (1 study), 4 years (1 study), and > 5 years (5 studies).

4. Pooled cumulative incidence in septic patients at longest follow-up was 3% (95% CI 0-5%; 10 studies) for #cvMI, 6% (95% CI 1–12%; 10 studies) for #stroke, 9% (95% CI 2–17%; 9 studies) for #heartfailure, and 10% (95% CI 5-15%; 18 studies) for composite #CVD.

5. #Sepsis was associated with ⬆️#cvMI (aHR 1.77 [95% CI 1.26–2.48]; low certainty), ⬆️#stroke (aHR 1.67 [95% CI 1.37–2.05]; low certainty), and ⬆️#heartfailure (aHR 1.65 [95% CI 1.46–1.86]; very low certainty) at longest follow-up c/w non-sepsis controls.

6. These associations were observed regardless of comparison with non-sepsis hospitalized or population-based controls.

7. Conclusion: #Sepsis may be an important risk factor for the development and pathogenesis of short- and long-term #CVD. This may represent a critical and underutilized opportunity for #cvPrev, which may also stretch beyond sepsis to a broad spectrum of critical illness.

8. Limitations: The evidence from our #metaanalysis was low-certainty, likely due to wide variation in population + outcome selection/timing and degree of adjustment for baseline #CVD RFs. Further studies with⬇️risk of bias and robust adjustment for underlying #CVD are needed.

10: Many thanks to my co-first author and intensivist extraordinaire @fedeangriman and the support of our mentor @PatrickLawlerMD . Thanks also to our incredible co-authors @LiorBibasMD , Emma Katz, Neill Adhikari, @LucasCGodoy @DrJohnCMarshall @BrunoFerreyro @DouglasLeeMD...

11. ...Robert Rosenson, Naveed Sattar, Subodh Verma, Augustin Toma, Marina Englesakis, @BarryBursteinMD @drmikefarkouh , Margaret Herridge, @denniskomd @damon_scales @michael_detsky and all of our mentors at @UofTCardio @UofT_DoM @uoftmedicine.