Repolarization alternans is a marker of increased vulnerability to fatal arrhythmias. At the tissue level, in unipolar electrograms (UEGs) recorded on the myocardium, repolarization alternans is often measured as an alternating change of the T-wave, so called T-wave alternans (TWA). During ischemia, UEG-TWA is used as a marker of cardiac instability and is considered as a key parameter to assess pharmacological strategies. However, during ischemia it is not clear whether UEG-TWA is a sign of repolarization alternans which may promote 2:1 conduction block, or whether it is induced by ongoing regional 2:1 conduction block. In this study, we first show in-vivo human data recorded during an ischemic event that suggest that 2:1 conduction block induces UEG-TWA beyond the region of 2:1 conduction block. We then develop an analytical forward model of the UEG by coupling an analytical description of the cardiac action potential with a theoretical expression of the UEG, where each UEG is the combination of a local and a remote component and noise. With this model, we were able to generate signals that closely resemble UEGs recorded in-vivo, with a maximum correlation ρ > 0.94. Finally, we interrogate the model and demonstrate that whenever 2:1 conduction block is present, UEG-TWA arises as a consequence of alternating imbalance of both the local and remote components of the UEG. The statistical significance of UEG-TWA depends on the interactions between local and remote dynamics and noise.We conclude that in an ischemic model, UEG-TWA is likely to be a sign of 2:1 conduction block, either proximal or distal from the recording site.