Morbid obesity in women on the rise: an observational, population-based study

The obesity epidemic in Australia is generally monitored by the proportion of the population whose body mass index (BMI) exceeds 30 kg/m² [1]. However, what is not as evident is the growing proportion of those who are morbidly obese. This issue is important as adverse health risks and healthcare expenditure amplify as the level of obesity increases [2]. Obesity-related health expenditure in Australia exceeded $8 billion in 2008, which included expenditure associated with metabolic disease, cardiac disease and surgical complications [3], but it is unclear how much of this expenditure related to different levels of obesity. Service providers are increasingly having to invest in equipment designed to accommodate the morbidly obese. It seems that both direct and indirect costs will amplify as the prevalence of morbid obesity increases.

In Australia, few, if any, studies have described the pattern of morbid obesity in the population. Thus, we aimed to document changes in the prevalence of morbid obesity (BMI ≥ 40.0 kg/m²) that have occurred over a decade among women living in south-eastern Australia.

For the period 1993–7, 34 (2.3%) women were identified as morbidly obese (class III obesity), 75 (5.0%) as class II obesity and 227 (15.2%) as class I obesity; there were 494 (33.1%) women classified as overweight; 635 (42.5%) as having normal weight and 29 (1.9%) as underweight (Table 1). A decade later, 2004–8, 44 (4.1%) women were morbidly obese (class III obesity), 81 (7.5%) were class II obesity, 191 (17.8%) were class I obesity, and 333 (31.0%) were overweight, 412 (38.3%) were of normal weight and 15 (1.4%) were underweight.

The odds for morbid obesity were 76% greater in 2004–8 compared to 1994–7 (OR = 1.76, 95%CI 1.12-2.78). This association was sustained after adjusting for SES (data not shown) and found to be consistent across SES categories.

The age-standardised prevalence for morbid obesity (class III obesity) was 2.5% for 1993–7 and 4.2% for 2004–7; prevalence figures for class II obesity were 5.6% and 7.3%, for class I obesity 15.1% and 17.6%, for overweight 31.8% and 30.8%, for normal weight 43.4% and 38.8%, and for underweight 1.6% and 1.3% (Figure 1). The SMR for morbid obesity (class III obesity) was 1.69 (95%CI 1.26-2.27).

We report that mean BMI has increased by 4.2% among women during the decade between 1993–7 and 2004–8. A comparison of the prevalence of morbid obesity in 2004–8 with 1993–7 generated a SMR for morbid obesity between 1.26 and 2.27, demonstrating a measurable increase over this ten year time period. Because of the approximate bell-shaped distribution of BMI in the population, a shift of the distribution to higher BMI resulted in a marked proportional increase in the prevalence of the more extreme BMI values. Disproportionate increases in the more extreme categories of BMI have been similarly reported for adults in the USA between 2000 and 2005 [7].

Our results indicate that the increase in the proportion of morbid obesity was not restricted to particular ages or different SES groups. Greater adiposity, and unhealthy lifestyle choices associated with increased risk for obesity, have been observed with increasing social disadvantage [8]. It is curious that our current study showed an increase in the prevalence of morbid obesity to have occurred across the SES spectrum. We may speculate that the increase in morbid obesity affects the population as a whole, thus shifting the entire population distribution toward greater risk for obesity-related disease. However, given the small number of morbidly obese women in our study, there may have been insufficient power to conclusively test for SES differences and it is plausible that a larger sample size could show inequity in the rise of morbid obesity between SES groups. Either way, our data suggest that population-wide approaches that change the underlying conditions in which behaviours occur could be warranted to combat rising rates of morbid obesity for all sub-groups of the population.

The strength of this study lies in the random nature of the study population and that the anthropometric measures used to calculate BMI were measured, rather than self-reported which would increase the potential for misclassification [9, 10]. However, it is possible that participation bias may have changed over time. It is not unreasonable that social stigma linked with obesity [11] may change as the obesity epidemic progresses, and we cannot exclude the possibility that this might have differentially impacted on participation at the two time periods. Moreover, limitations of BMI as an indicator of adiposity have previously been demonstrated, especially for individuals with muscular body builds and the elderly [1], and this was not considered in the analyses.

Our study findings highlight the growing problem of morbid obesity associated with a population-wide increase in BMI. Recently, there has been increased advocacy in Australia for raising awareness of nutrition-related diseases and integrating nutrition into medical curricula [12]. In further recognition of issues surrounding the clinical management of morbidly obese patients, a Morbid Obesity model of care was developed in Western Australia in 2007 [13]. It should be clear, however, that unless the modern obesogenic environment is addressed, obesity-related health problems and healthcare costs are likely to escalate. In 2009, a parliamentary committee developed a series of recommendations designed to combat the rising rates of obesity in Australia [3]. Recommendations included development of the built environment to promote physically active lifestyles, education to encourage healthy eating and to increase the consumption of fresh fruit and vegetables, reformulation of commercially-available foods, more effective food labelling to assist healthy consumer choices and the development of a registry of bariatric surgery. While surgery and other gastrointestinal interventions may be considered appropriate for individuals with severe obesity and co-morbid disease, such as type 2 diabetes [5], such treatment requires a thorough multidisciplinary team assessment of peri-operative risk and possible long-term complications [14] and is unlikely to impact the rising prevalence of morbid obesity at a population level. The challenge is to identify and implement effective strategies that will shift the population distribution towards lower BMI, if the disproportionate increase observed in the extreme levels of obesity is to be curbed.

We conclude that over a period of a decade, there has been an increase in mean BMI among women residing in south-eastern Australia, resulting in a measurable rise in the prevalence of morbid obesity.