Hundreds of volatile organic compounds (VOCs) are released from human body fluids. Some of them are produced by endogenous metabolic processes in and on the body, and others are derived from the environment. Expressions of some endogenous VOCs can be affected by pathophysiological changes, and several disease-specific volatile biomarkers have been identified and used as diagnostic aids. Monitoring volatile disease markers is attractive since the procedure can be performed in a noninvasive manner with little or no exposure to biohazardous body fluids. Although many VOCs have been claimed as potential biomarkers, only a few compounds have been consistently demonstrated and approved for clinical applications. This is mainly because (1) many of the putative markers are present in the environment as well as in the body and their levels in the environment are often higher than those in the body, (2) there are a large individual variation in the concentrations of biomarkers within diseased and/or healthy subjects, and (3) the origin and biosynthetic pathway of the claimed biomarkers have been frequently neglected. Unfortunately, these aspects have often been ignored in many studies. Here, we review a number of publications that have identified volatile disease biomarkers in breath, argue that many of these have not demonstrated to actually underlie the differences in volatile profiles between diseased patients and healthy subjects, speculate on the reasons for this lack of success, and suggest potential approaches that may provide a better chance of identifying disease biomarkers.