Background: Average serum matrix metalloproteinase (MMP) concentrations in patients with acute stroke have shown to be varying across studies. Possibly, next to true effects, other factors may influence MMP levels. The aim of this study was to investigate the dynamics of these enzymes in repeated measurements in the acute post-stroke period, in respect to different stroke etiologies, and highlight potential sources for variability. Methods: Serum in 233 patients with acute ischemic or hemorrhagic stroke (stroke cohort; SC) was ascertained within 24 h after onset and then 1, 3 and 7 days thereafter. One hundred five controls (control cohort; Co) were recruited. Multi-variable adjustment was carried out using salient extraneous covariates including stroke etiology, clinical severity and lesion size next to a set of routine laboratory parameters. Results: Unadjusted SC MMP-2 concentrations are significantly lower (SC 165.4, 95% CI 158.5-172.4; Co 203.7 ng/ml, 95% CI 190.7-216.5; p < 0.001) and MMP-9 concentrations significantly higher than in controls (SC 608.5 ng/ml, 95% CI 555.3-661.8; Co 475.6 ng/ml, 95% CI 413.6-537.6; p < 0.001). Adjustment mitigates associations between MMP concentrations and stroke etiology, clinical severity, lesion size or differences in temporal profile shown present without adjustment. Salient covariates absorb much of the effect: age, leukocyte count and albumin concentrations are associated significantly with MMP-2 concentrations; only leukocyte count is significantly associated with MMP-9. Conclusions: Concentrations of MMP-2 and MMP-9 in serum in humans measured after acute stroke are potentially influenced by extraneous covariates rather than being directly associated with characteristics of the underlying stroke.