We found a significant association between self-reported lack of sleep and IGT in men that remained after adjustment for potential confounders such as obesity and health behaviours. No such association was found in women. Furthermore, a significant association was found between low vitality and IGT in both men and women. While the association was more robust in women, it was lost in men when other patient-related factors were considered. The study also confirmed that both men and women with IGT have a more serious cardiovascular disease risk factor profile than participants with NGT. A worse risk factor profile was also revealed in both men and women with IFG as compared to NGT, however, we found no associations with lack of sleep or low vitality.
In support of our findings, the Quebec Family Study  found that both short and long total sleeping time predicted type 2 diabetes/IGT in adult men and women. Clinical effects of sleep deprivation are associated with common symptoms such as sleepiness, increased fatigue, and low motivation , a finding that is in agreement with the results of this study. Furthermore, voluntary sleep reduction is increasingly becoming a way of life – a habit related to the modern 24 hour society , and studies show that nearly 30% of the middle-aged population reported sleeping less than 6.5 h per night . As a consequence, reports of day-time sleepiness have become more frequent during recent years . The cause of sleep-loss is multi factorial, although obesity is considered an important risk factor for obstructive sleep apnea (OSA) . Accordingly, in the FIN-D2D survey , middle-aged men with sleep-disordered-breathing (SDB) had an increased incidence of type 2 diabetes and abnormal glucose tolerance. However, no corresponding association has been found in women in either the FIN-D2D survey  or the study of women in Gothenburg .
The significant association found between level of vitality and IGT was robust in women, but was diluted and lost in men when successively accounting for other patient-related factors. As fatigue is one of the more common presenting symptoms of type 2 diabetes, this is interesting. Physical activity came out as an explanatory factor in men, but an interaction should be expected in both genders by virtue of its strong association with insulin resistance and consequently the development of both IGT and type 2 diabetes . Since physical activity is known to improve mood, and a good mood most likely increase the motivation for physical activity it may thus also be associated with the vitality factor .
As shown in a prospective study, psychological distress such as fatigue, anxiety and insomnia increases the risk for prediabetes and type 2 diabetes in Swedish middle-aged men . There is a common belief that day-time sleepiness is normal or related to poor life style or laziness, especially if it interferes with daily functions . However, OSA often characterized by daytime sleepiness, and diabetes share common mechanisms including age and obesity, but the direction of causality may go both ways . Consequently, OSA should also be considered when these symptoms are investigated. The underlying mechanism also involving genetics, for example a mutation in the melatonin receptor 1B, should also be considered .
Our study question on sleep function did not account for sleep duration or possible SDB. However, the difference between men and women in the association between lack of sleep and IGT is consistent with the findings in the FIN-D2D study , and might thus be explained by the fact that men are more exposed to OSA, or more susceptible to the effects of OSA. Nevertheless, our findings were not diluted when BMI was adjusted for. Unfortunately, we did not measure breathing pattern during sleep directly or by questionnaire. In a previous Swedish study, men were also more susceptible to psychological distress in the association with prediabetes than women . We found no corresponding pattern in the association with low vitality. This may probably be explained by low vitality being derived from other mechanisms than psychological distress.
Strengths and limitations of the study
This study is based on a large, random population sample with a high participation rate, making the results generalizable to this and other similar populations. A further strength of this study is the enrolment of both men and women over a wide age-band where strategies of diabetes prevention are important. The prevalence of IGT was also congruent with other studies from Sweden .
An OGTT was performed in each participant without a known diagnosis of diabetes. According to recommendations from WHO, an OGTT should be performed to diagnose IGT as it is characterized by postprandial hyperglycemia and separate from IFG that is characterized by fasting hyperglycemia. This procedure to identify IGT is supported by experiences from other population studies . The questionnaires were completed before the information of the results of the OGTT was available, and therefore none of the participants were aware of their potential diagnosis of IGT when answering the questionnaire. In this study we used self-reported information on physical activity on physical activity in leisure time, (LTPA). However, this question has shown good validity when recently compared to objectively measured total physical activity during 24 hour . Limitations of the study comes from it’s cross-sectional design, making it impossible to decide on causality in associations. Information on sleep function and vitality was self-reported and not based on direct measurements; however, these questions have been used and validated in other studies with reliable findings. Thus, non-restorative sleep was explored in the Minimal Insomnia Symptom Scale (MISS) , and our question on vitality was included in a validation of the SF-36 questionnaire . The question on self-experienced lack of sleep in this study is a general perception on the sleep quality covering both sleep duration and sleep quality as compared to instruments that evaluate more specific domains of sleep quality like the Pittsburg Slep Quality Index . It still correlates well with other characteristics of sleep function . Finally, the protocol did not comprise any measurements on OSA and HbA1c was not measured in the complete sample, and thus these important factors could not be accounted for.