How many people have eating disorders?

It's important to distinguish between incidence and prevalence here. Incidence is defined as the number of new cases of a condition over a particular period of time (usually a year), whilst prevalence is the number of individuals in a given population who have the condition at any one point in time. Note that prevalence does not tell you how many people have ever suffered with the condition.

So, here are prevalence and incidence data for the eating disorders as far as I have been able to discover. Note that all the data are based on studies from Europe and North America.

Anorexia Nervosa:
Widely differing data are available regarding the incidence of AN. For example, a 1985 study in the Netherlands found a rate of 18.4 per 100,000, whilst an American study conducted at around the same time found an incidence of 149.5 per 100,000 (Lucas et al; 1991). Prevalence of AN is about 0.3% in young women (Hoek and van Hoeken 2003).

Consensus suggests that the disorder is probably not increasing in frequency (Hoek and van Hoeken 2003), although this might appear to be the case due to better clinical identification.
The natural course of AN is hard to determine due to measurement problems and drop out, but is likely to be very variable, depending on treatment and other factors. However, meta-analyses suggest that 50-70% achieve intermediate or good long term outcome (Pike 1998)

Bulimia Nervosa:
The incidence of BN is around 12/100,000, measured in the UK (Turnbull et al, 1996), USA (Soundy et al, 1995) or the Netherlands (Hoek et al, 1995). Prevalence differs widely depending on age and gender but appears to be around 1% in young women and about 0.1% in young men (Hoek and van Hoeken 2003). Some authors suggest that it may be becoming more common (Turnbull, Ward et al. 1996).

With regard to the course of the disorder, Fairburn and colleagues study of a community sample of UK women with BN is particularly informative. Fifteen months after diagnosis, half to two thirds of these women still had some kind of eating disorder. Each year that they were followed (up to a total of 5 years), about one third of the those in remission relapsed, whilst about one third of those who remained ill entered remission (Fairburn, Cooper et al. 2000).

Eating Disorder Not Otherwise Specified (EDNOS):
Few data are available on this diagnostic group, despite evidence suggesting that it is the most common diagnosis of individuals presenting to eating disorder services (Williamson, Gleaves et al. 1992; Turner and Bryant-Waugh 2004).

References
Fairburn, C., Z. Cooper, et al. (2000). "The natural course of bulimia nervosa and binge eating disorder in young women." Archives of General Psychiatry 57: 659-665.
Hoek, H. and D. van Hoeken (2003). "Review of the prevalence and incidence of eating disorders." International Journal of Eating Disorders 34: 383-96.
Pike, K. (1998). "Long-term course of anorexia nervosa: response, relapse, remission and recovery." Clinical Psychology Review 18(4): 447-475.
Turnbull, S., A. Ward, et al. (1996). "The demand for eating disorder care: an epidemiological study using the General Practice Research Database." British Journal of Psychiatry 169: 705-712.
Turner, H. and R. Bryant-Waugh (2004). "Eating Disorder Not Otherwise Specified (EDNOS): Profiles of Clients Presenting at a Community Eating Disorder Service." European Eating Disorder Review 12: 18-26.
Williamson, D., D. Gleaves, et al. (1992). "Empirical classification of Eating Disorder Not Otherwise Specified: Support for DSM-IV changes." Journal of Psychopathology and Behavioural Assessment 14: 201-216.