Selective COX-2 inhibitors (coxibs) increase the incidence of cardio- and cerebro-vascular events. Complete disruption of the murine gene encoding COX-2 (Ptgs2) leads to renal developmental problems, as well as female reproductive anomalies and patent ductus arteriosus of variable penetrance in newborns, thus rendering this genetic approach difficult to compare with coxib administration. Here, we created hypomorphic Ptgs2 (COX-2^Neo/Neo) mice in which COX-2 expression is suppressed to an extent similar to that achieved with coxibs, but not eliminated, in an attempt to circumvent these difficulties. In LPS-challenged macrophages and cytokine-stimulated endothelial cells obtained from COX-2^Neo/Neo mice, COX-2 expression was reduced 70-90 % and these mice developed a mild renal phenotype compared to COX-2 mice possessing an active site mutation (COX-2^Y385F/Y385F), with minimal signs of renal dysfunction as measured by FITC-inulin clearance and blood urea nitrogen. These COX-2 knockdown mice displayed an increased propensity for thrombogenesis compared to their wild-type (COX-2^+/+) littermates observed by intravital microscopy in cremaster muscle arterioles upon ferric chloride challenge. Measurement of urinary prostanoid metabolites indicated that COX-2^Neo/Neo mice produced 50 % less prostacyclin but similar levels of PGE₂ and thromboxane compared to COX-2^+/+ mice in the absence of any blood pressure and ex vivo platelet aggregation abnormalities. COX-2^Neo/Neo mice, therefore, provide a genetic surrogate of coxib therapy with disrupted prostacyclin biosynthesis that predisposes to induced arterial thrombosis.