A recent systematic review, published in the BMJ, has concluded that in countries where data is available the incidence of diabetes has been stable or falling since 2006.

The authors performed a systematic review of studies reporting trends of diabetes incidence. The PRISMA checklist for systematic reviews was used to search multiple databases including Medline, Embase and CINAHL for suitable studies from 1980 to 2017. Studies of open population-based cohorts, diabetes registries and administrative and insurance databases that reported secular trends of diabetes were included. Eligible studies had to report total diabetes or type 2 diabetes incidence in two or more time periods.

Of 22,833 screened abstracts, 47 studies were included in the analysis. The countries represented in the studies were: Canada, China, Denmark, Finland, Germany, Hong Kong, Iceland, Israel, Italy, Korea, Mauritius, Nauru, Netherlands, Norway, Portugal, Scotland, Sweden, Switzerland, Taiwan, UK and USA.

Trends in incidence of diabetes

1960-89

• 36% of the populations studied had increasing trends

• 55% had stable trends

• 9% had decreasing trends

1990-2005

• 66% had increasing trends

• 32% had stable trends

• 2% had decreasing trends

2006-2014

• 33% had increasing trends

• 30% had stable trends

• 36% had declining trends

The incidence of diagnosed diabetes increased in most populations up to the early 2000s. The majority of the populations studied showed stable or reducing incidence of diabetes from the mid 2000s.

Countries reporting a decrease in incidence over this period included the USA, Israel, Switzerland, Hong Kong, Sweden and Korea. Stable incidence was reported from the UK, Canada, Italy, Scotland, Norway, and the USA (non-hispanic white).

Increasing incidence was only found in Portugal, Denmark and Germany.

The authors acknowledge that the systematic review could not adjust for different diagnostic thresholds for diabetes, and sometimes for diabetes type, although type 2 diabetes is always the most common. The data were limited in low- and middle-income countries.

Even with these caveats, this paper has interesting messages about the paridigm of the diabetes epidemic. The authors point out that prevalence of type 2 diabetes may increase because of improved survival, and that the simultaneous decline in incidence could be a reduction in the pool of undiagnosed diabetes patients, together with improving health awareness.

They cite increases in exercise from interventions at schools and workplaces, reductions in the consumption of sweetened beverages and active diabetes prevention programmes as examples of interventions that may have reduced the incidence of diabetes.

We have tended to feel that the rising trends in childhood obesity, and the traditional statistics of the large pool of undiagnosed diabetes have been part of an uncontrollable diabetes epidemic.

This paper may suggest that there are some grounds for cautious optimism about the future, especially in conjunction with the accumulating evidence for changes in lifestyle preventing, and even reversing, diabetes.

Dr Matthew Lockyer