Anybody read this article yet?  The abstract on Medline was very promising.

Pettitt DJ, Forman MR, Hanson RL, Knowler WC, Bennett PH,  Breastfeeding and
incidence of non-insulin-dependent diabetes mellitus in Pima Indians. Lancet
1997 Jul 19;350(9072):166-168

BACKGROUND: Early exposure to cow's milk has been implicated in the occurrence
of insulin-dependent diabetes mellitus but there is little information about
infant-feeding practices and subsequent non-insulin-dependent diabetes mellitus
(NIDDM). We examined the association between breastfeeding and NIDDM in a
population with a high prevalence of this disorder, the Pima Indians. METHODS:
Glucose-tolerance status was obtained from a 75 g oral glucose-tolerance test. A
standard questionnaire given to mothers was used to classify infant-feeding
practices for the first 2 months of life into three groups; exclusively
breastfed, some breastfeeding, or exclusively bottlefed. The association between
the three infant-feeding groups and NIDDM was analysed by multiple logistic
regression. FINDINGS: Data were available for 720 Pima Indians aged between 10
and 39 years. 325 people who were exclusively bottlefed had significantly higher
age-adjusted and sex-adjusted mean relative weights (146%) than 144 people who
were exclusively breastfed (140%) or 251 people who had some breastfeeding
(139%) (p = 0.019). People who were exclusively breastfed had significantly
lower rates of NIDDM than those who were exclusively bottlefed in all age-groups
(age 10-19, 0 of 56 vs 6 [3.6%] of 165; age 20-29, 5 [8.6%] of 58 vs 17 [14.7%]
of 116]; age 30-39, 6 [20.0%] of 30 vs 13 [29.6%] of 44). The odds ratio for
NIDDM in exclusively breastfed people, compared with those exclusively
bottlefed, was 0.41 (95% CI 0.18-0.93) adjusted for age, sex, birthdate,
parental diabetes, and birthweight. INTERPRETATION: Exclusive breastfeeding for
the first 2 months of life is associated with a significantly lower rate of
NIDDM in Pima indians. The increase in prevalence of diabetes in some
populations may be due to the concomitant decrease in breastfeeding.