Clinical management of heart failure requires reliable and accurate tools to categorize patients on the basis of underlying causes, associated risks, treatment responses, and outcomes. In recent years, myocardial transcriptional profiling has been used extensively to this end. Most prior transcriptional profiling studies in heart failure, however, examined only cardiac mRNAs or microRNAs (miRNAs), which together constitute only a small portion (≈1%) of the human genome. Because a large proportion of the genome is transcribed as long noncoding RNA (lncRNA), we profiled lncRNAs in parallel with mRNAs and miRNAs to provide a complete picture of the human cardiac transcriptome landscape in heart failure. Exploiting next-generation sequencing technology, we conducted a comprehensive cardiac transcriptome profiling study encompassing mRNAs, miRNAs, and lncRNAs in human left ventricular samples from heart failure patients before and after mechanical circulatory support with a left ventricular assist device and in left ventricular samples from nonfailing donors. We demonstrate here that the expression signature of cardiac lncRNAs compared with those of mRNAs or miRNAs provides exquisite sensitivity to discriminate cardiomyopathic samples of ischemic and nonischemic origins and to distinguish heart failure samples before and after left ventricular assist device support, revealing the potential of using lncRNA profiles as a reliable biomarker to differentiate between cardiac disease states. We further provide evidence to show that cardiac lncRNAs likely function through regulating the expression of nearby (cis) rather than distant (trans) genes. Taken together, these studies suggest an important role for lncRNAs in the pathogenesis of heart failure and in reverse remodeling observed after ventricular unloading with mechanical support.