BackgroundPercutaneous coronary intervention (PCI) aims to relieve coronary vessel obstruction to improve coronary blood flow and relieve cardiac ischaemia. Some studies suggest little improvement in outcomes compared with medication alone, whereas others suggest that PCI is beneficial if physiologically guided. Few studies have assessed whether PCI improves coronary blood flow and how any change in flow is related to stenosis severity. We sought to determine whether fractional flow reserve (FFR), a pressure-only physiological index of stenosis severity and surrogate for ischaemia, could predict the improvement in coronary flow produced by PCI. MethodsWe assessed 75 stenoses in 67 patients undergoing elective PCI. We conducted simultaneous transtenotic pressure and Doppler flow velocity measurements during adenosine-mediated hyperaemia, both before and after PCI. Flow velocity was calculated over the whole cardiac cycle. Change in flow was stratified by lesion severity, with particular interest to flow changes in stenoses with FFR more than 0·80, which is typically classed as non-ischaemic. FindingsHyperaemic flow velocity was significantly higher after PCI than before PCI (mean 0·30 m/s [SE 0·02] vs 0·51 [0·03], p<0·001) across all stenoses. Both the absolute and relative increase in blood flow after PCI was strongly correlated to the pre-PCI FFR measurement (r=−0·65, p<0·001), with a small non-significant but detectable increase in flow when FFR was borderline (FFR 0·71–0·80: mean change from pre-PCI [Δ] 0·11 m/s [0·07], p=0·14). The greatest incremental increase was seen when FFR in stenoses was less than 0·60 (Δ 0·41 [0·05], p<0·001). By contrast, when pre-PCI FFR was more than 0·80, PCI produced little improvement in flow (0·05 m/s [0·02]) and this change was significantly less than when stenoses were classed before PCI as flow limiting (p<0·001). InterpretationBy removing epicardial stenoses, PCI improves hyperaemic coronary blood flow, and this increase is strongly related to the physiological significance of the stenosis before PCI. Only stenoses defined as severe by FFR, with values consistent with ischaemia, had a significant rise in flow velocity after PCI. Stenoses deemed non-significant by FFR had comparatively little improvement in flow. These findings provide a mechanistic explanation for the improved outcomes and symptomatic improvement after physiologically guided stenting, and strongly support restricting PCI to patients with truly flow-limiting coronary stenoses. FundingUK Medical Research Council.