Ten studies were included (n=2,363); four RCTs (n=198) and six controlled cohort studies (n=2,165). One RCT scored two points and three scored three points (out of six points) on the Jadad scale. All RCTs used intention to treat analysis. No studies reported blinded outcomes assessment and all were underpowered. Cointerventions were poorly described in the cohort studies.
Hospital mortality: RCT evidence.
There was a significantly lower hospital mortality rate in the intervention group (OR 0.18, 95% CI: 0.06, 0.57, p=0.003; ARR 15%, NNT 7, 4 RCTs) without evidence of significant heterogeneity. The IABP group had significantly shorter hospital stay and cardiopulmonary bypass time and a lower rate of low cardiac output syndrome. No cases of IABP-related mortality or aortic dissection were reported. Leg ischaemia occurred in five patients in the intervention group (n=99) and five controls who received IABP (n=39). Leg infection and bleeding at the IABP insertion site were uncommon.
Hospital mortality: RCT and cohort evidence.
When all ten studies were pooled, there was a significantly lower hospital mortality rate in the intervention group (OR 0.41, 95% CI: 0.21, 0.82, p=0.01; ARR 6%, 95% CI: 2, 10, p=0.007; NNT 17, 95% CI: 10, 50), with significant heterogeneity (p=0.03). There was no evidence of publication bias on the funnel plot. There was no IABP-related mortality. Preoperative IAPB was associated with limb ischaemia or haematoma at the insertion site in 3.7% (13/349) of cases. Most complications resolved with discontinuation of IABP.