Background Although B-type natriuretic peptide (BNP) is used as complimentary diagnostic tool in patients with unknown thoracic disorders, many other factors appear to trigger its release. In particular, it remains unresolved to what extent cellular stretch or wall stress of the whole heart contributes to enhanced serum BNP concentration. Wall stress cannot be determined directly, but has to be calculated from wall volume, cavity volume and intraventricular pressure of the heart. The hypothesis was, therefore, addressed that wall stress as determined by cardiac magnetic resonance imaging (CMR) is the major determinant of serum BNP in patients with a varying degree of left ventricular dilatation or dysfunction (LVD). Methods A thick-walled sphere model based on volumetric analysis of the LV using CMR was compared with an echocardiography-based approach to calculate LV wall stress in 39 patients with LVD and 21 controls. Serum BNP was used as in vivo marker of a putatively raised wall stress. Nomograms of isostress lines were established to assess the extent of load reduction that is necessary to restore normal wall stress and related biochemical events. Results Both enddiastolic and endsystolic LV wall stress were correlated with the enddiastolic LV volume (r = 0.54, P < 0.001; r = 0.81, P < 0.001). LV enddiastolic wall stress was related to pulmonary pressure (capillary: r = 0.69, P < 0.001; artery: r = 0.67, P < 0.001). Although LV growth was correlated with the enddiastolic and endsystolic volume (r = 0.73, P < 0.001; r = 0.70, P < 0.001), patients with LVD exhibited increased LV wall stress indicating an inadequately enhanced LV growth. Both enddiastolic (P < 0.05) and endsystolic (P < 0.01) wall stress were increased in patients with increased BNP. In turn, BNP concentration was elevated in individuals with increased enddiastolic wall stress (>8 kPa: 587 +/- 648 pg/ml, P < 0.05; >12 kPa: 715 +/- 661 pg/ml, P < 0.001; normal < or =4 kPa: 124 +/- 203 pg/ml). Analysis of variance revealed LV enddiastolic wall stress as the only independent hemodynamic parameter influencing BNP (P < 0.01). Using nomograms with "isostress" curves, the extent of load reduction required for restoring normal LV wall stress was assessed. Compared with the CMR-based volumetric analysis for wall stress calculation, the echocardiography based approach underestimated LV wall stress particularly of dilated hearts. Conclusions In patients with LVD, serum BNP was increased over the whole range of stress values which were the only hemodynamic predictors. Cellular stretch appears to be a major trigger for BNP release. Biochemical mechanisms need to be explored which appear to operate over this wide range of wall stress values. It is concluded that the diagnostic use of BNP should primarily be directed to assess ventricular wall stress rather than the extent of functional ventricular impairment in LVD.