Atrial fibrillation (AF) increases stroke risk and is prevalent in heart failure (HF)1-4. Alternans, a beat-to-beat oscillation in the shape of the atrial action potential and/or Ca2+ transient, has been implicated in AF initiation1 however the aetiology of alternans is not fully understood. Atrial contraction and relaxation is brought about by the rise and fall in cytosolic Ca2+ during the systolic Ca2+ transient. The machinery responsible for producing this rise and fall of Ca2+ is concentrated around a network of membrane invaginations or transverse (t)-tubules which ensure Ca2+ release is triggered at sites throughout the cell. T-tubules are reduced in AF and almost entirely lost in HF resulting in a loss of triggered Ca2+ release sites5. Evidence suggests changes in intracellular Ca2+ cycling play an important role in cellular alternans. We speculated that t-tubule loss may predispose to Ca2+ dependant alternans which may contribute to the increase in the propensity for AF in HF. To determine if alternans could be induced more easily in HF and if this was due to t-tubule loss. HF was induced by pacing the right ventricular of adult (~18 months) female sheep at 210 beats per minute. Pacemaker implantation was performed under anaesthesia with 2-5% (v/v) isoflurane in oxygen (4.5 L.min-1). Left atrial cardiac myocytes were isolated from control and HF sheep. The HF induced t-tubule loss was mimicked in control cells by formamide induced detubulation (verified confocally). Fluo-5F loaded myocytes were incrementally paced between 1 and 3 Hz under perforated patch current clamp control to induce alternans. The lowest frequency at which alternans was detected was deemed the threshold. Values are means ± S.E.M and compared by one-way ANOVA. The threshold for Ca2+ alternans was decreased in HF vs. control (1.25 ± 0.18 vs. 2.40 ± 0.16 Hz; p<0.05, n=13-44) and for detubulated cells vs. control (1.75 ± 0.34 vs. 2.40 ± 0.16 Hz; p<0.05, n=9-44). There was no significant difference between HF and detubulated cells. T-tubule loss, either during HF or following detubulation with formamide had no effect on the threshold for action potential alternans (2.41 ± 0.10 Hz control vs., 2.00 ± 0.20 Hz in HF and 2.19 ± 0.24 Hz following detubulation; n=9-44/6-28 cells/animals). Our data suggests alternans occurs more readily in atrial cells from animals at the point of HF. However further work is required to establish if this plays a role in the increased prevalence of AF in HF. Our data suggests that t-tubule loss contributes to the increased susceptibility to alternans. Further work is required to determine whether other key components of the intracellular Ca2+ cycling are mechanistically involved in the earlier onset of alternans in HF and how these components alter with t-tubule loss.