Conference Coverage

Mass Diabetes Screening Doesn't Change Death Rates

View on the News


The report of the ADDITION-Cambridge trial raises many questions. Surely, the epidemic of diabetes and the established benefits of early aggressive therapy beg the question of early aggressive screening. The study fails to demonstrate any benefit of screening in 10 year diabetes-, cancer-, or cardiovascular-related mortality in those persons at “high risk” for diabetes, compared with unscreened individuals.

However, two elements of the trial design deserve a closer look: the definition of “high risk” and the sole emphasis on mortality as an outcome measure.

Patients in this study were identified as being at high risk for undiagnosed diabetes on the basis of data related to age, sex, body mass index, and steroid and antihypertensive medication. Missing from this category would appear to be family history of diabetes, evidence of insulin resistance such as diabetic dyslipidemia, history of coronary artery or macrovascular disease and other associated disorders such as polycystic ovary syndrome, nonalcoholic fatty liver disease, microalbuminuria, or evidence of peripheral neuropathy. Furthermore, certain ethnic and geographic populations are at much higher risk for diabetes than others. The data from this study cannot be accurately applied to those populations. The benefit of screening would certainly be amplified by expanding the high-risk definition to include these common and diabetes related disorders. My personal clinical experience includes a high degree of detection of both impaired glucose tolerance and previously undiagnosed diabetes in patients with these disorders, particularly those with a history of coronary artery disease or with a family history of diabetes.

Dr. Paul S. Jelllinger

Regarding the trial’s primary outcome measure, the sole focus on mortality – rather than the substantial morbidity associated with the microvascular complications of diabetes – is an important modifying factor. Since microvascular complications have been shown to be closely related to glycemic control, screening with the intent to reduce microvascular disease may have a significant quality-of-life and cost-effectiveness benefit. Data relating to screening and the ability to impact microvascular disease is sorely needed.

Each physician needs to assess their patient’s risk for diabetes and the advisability to screen, not necessarily based on broad scaled population data, but on those risks inherent in that individual. An appreciation of the full array of disorders that place any individual patient at risk for diabetes will surely improve the detection rate of both prediabetes and diabetes. Both dysglycemic states share microvascular and macrovascular complications not necessarily reflected only in mortality rates.

PAUL S. JELLINGER., M.D., is a clinical endocrinologist in Hollywood, Fla,. and medical editor of CLINICAL ENDOCRINOLOGY NEWS. He has no relevant disclosures.



BERLIN – A population-based screening project for type 2 diabetes did not improve 10-year mortality, a very large British study has determined.

When compared to best-practice identification of individual diabetes patients, general screening offered no benefits in all-cause mortality or in the specific outcomes of cardiovascular, cancer, or diabetes-related deaths Dr. Rebecca K. Simmons reported Oct. 4 at the annual meeting of the European Association for the Study of Diabetes (EASD).

Dr. Rebecca K. Simmons

But the study did show that individuals at risk for the disease who refused the screening were significantly more likely to die over the follow-up period, said Dr. Simmons of the University of Cambridge (England).

"The take-home message here is that the benefits of population-based screening have been overestimated and are probably restricted to those who are at risk, who attend the screen, and who receive earlier treatment because of it," she said.

The ADDITION-Cambridge study involved 32 general medical practices in England and more than 20,000 patients who were at risk of developing type 2 diabetes. The practices were randomized to conduct either large-scale screening of all at-risk patients (27) or to no screening (5). The control practices continued to diagnose patients according to best practices.

Practices in the intervention group sent invitations to those who had been identified as high risk. The stepwise process included capillary blood glucose and hemoglobin A1c tests, and a confirmatory oral glucose tolerance test. Those who had the OGTT also had a cardiovascular risk factor assessment that included body mass index, waist circumference, blood pressure, and lipids. Those who did not attend the screen after a reminder were considered to be nonattenders.

Patients diagnosed in either practice group were managed according to a national consensus guideline of lifestyle education and modification, medical management, and blood pressure and cholesterol control.

All participants – whether screened or not – were followed for 10 years in the National Office of Statistics. The study accumulated 184,000 person-years of follow-up data.

Of the 16,000 screen-eligible patients, 15,000 were invited for screening and 11,737 (73%) attended the first stage. There were 466 new diagnoses of type 2 diabetes (3% of the screen-eligible group). There were 4,137 patients in the no-screening group, who served as controls.

Patients had a mean age of 58 years; most (64%) were male. The mean BMI was 30.5 kg/m2. About half were taking an antihypertensive, and a small number (4%-5%) were taking steroids.

At 10 years, there were 1,532 deaths in the screening group (10.5 per 1,000 person-years), and 377 in the control group (9.9 per 1,000 person-years) – a difference that was not significant.

The screening group had 482 cardiovascular deaths (3.3 per 1,000 person-years), compared with 124 in the control group (3.3 per 1,000 person-years). Cancer was the cause of death in 697 patients in the screening group and in 169 in the control group (4.8 and 4.4 per 1,000 person-years, respectively). Diabetes-related deaths occurred in 75 of the screening group and in 16 of the control group (0.5 and 0.4 per 1,000 person-years).

Deaths due to other causes occurred in 353 patients in the screening group (2.4 per 1,000 person-years) and 84 in the control group (2.2 per 1,000 person-years).

None of these findings were statistically significant, Dr. Simmons said.

However, a significant difference did emerge when the invited group was broken down into attenders and nonattenders. In this analysis, those who refused screening were 62% more likely to die within 10 years, and those who attended were 17% less likely to die, than were the control patients.

Most of those who refused screening were younger, obese men, who were less likely to be taking antihypertensive medications. Considered along with the significant increase in the risk of death in this group, targeted screening makes more sense than a general approach, Dr. Simmons suggested.

The results of ADDITION-Cambridge were published online to coincide with the EASD meeting (Lancet 2012 Oct. 4 [doi: 10.1016/S0140-6736(12)61422-6]).

"It’s probably best to focus on identifying those who have diagnosable disease and treating them according to our best practices," she said.

The study was funded by the Wellcome Trust, the Medical Research Council, the National Health Service, and the National Institute for Health Research. Dr. Simmons had no financial disclosures.

Recommended Reading

Tomatoes Boosted Low HDL in Small Study
MDedge Cardiology
Lipid Abnormalities Common in Granuloma Annulare
MDedge Cardiology
Gastric Bypass Benefits Persist at 6 Years' Follow-Up
MDedge Cardiology
Severe Psoriasis Linked to Doubled Diabetes Risk
MDedge Cardiology
Diabetes Diagnosed at Acute MI Hits Women Harder
MDedge Cardiology