Background: Although sleep deprivation and shift-work are associated with risk for metabolic syndrome (MS), it is less evident whether small shifts in sleep duration (operationalized here as the difference between weekend (WE) and weeknight (WN) sleep duration) also predict risk for MS independently of average sleep duration. The purpose of the present study was to examine whether a difference in WE and WN sleep duration was found among adults across a wide age range, whether it predicted number of symptoms of MS after controlling for estimated sleep duration, whether the difference in sleep duration was moderated by age, and to test indirect effects via the specific diagnostic criteria for MS (blood pressure [BP], high-density lipoprotein, triglycerides, glucose, and waist circumference). Methods: The study reanalyzed archived data. Participants were business leaders attending workshops in 2007-2009 who self-reported WE and WN sleep duration, and provided physiological measures collected by health care professionals. Results: Shifts in sleep duration from WE to WN were reported throughout this sample. The difference predicted number of symptoms of MS independently of overall sleep duration and relevant covariates (age, gender, minority ethnicity, history of smoking, diet, exercise and body mass index). The association was not moderated by age, and included a significant indirect effect by way of changes in glucose level. Conclusion: Small inconsistencies in WE to WN sleep duration predicted risk for MS independently of average sleep duration and age. This association may be indirect via changes in glucose levels.