Multi-Drug-Resistant strains of Mycobacterium tuberculosis (MDR-TB) are a serious obstacle to global TB eradication. While most MDR-TB strains are infrequently transmitted, a few cause large transmission clusters that contribute substantially to local MDR-TB burdens. Here we examine whether the known mutations in these strains can explain their success. Drug resistance mutations differ in fitness costs and strains can also acquire compensatory mutations (CM) to restore fitness, but some highly transmitted MDR strains have no CM. The acquisition of resistance mutations that maintain high transmissibility seems to occur by chance and are more likely in strains that are intrinsically highly transmitted and cause many cases. Modern Beijing lineage strains have caused several large outbreaks, but MDR outbreaks are also caused by ancient Beijing and lineage 4 strains, suggesting the lineage is less important than the characteristics of the individual strain. The development of fluoroquinolone resistance appears to represent another level of selection, in which strains must surmount unknown fitness costs of gyrA mutations. The genetic determinants of high transmission are poorly defined but may involve genes encoding proteins involved in molybdenum acquisition and the Esx systems. In addition, strains eliciting lower cytokine responses and producing more caseating granulomas may have advantages for transmission. Successful MDR/XDR strains generally evolve from highly transmitted drug sensitive parent strains due to selection pressures from deficiencies in local TB control programs. Until TB incidence is considerably reduced, there will likely be highly transmitted strains that develop resistance to any new antibiotic.