Turchin and his colleagues analyzed data from 4,368 non-critical hospital admissions. Of these, 2,582 were people with diabetes -- both type 1 and type 2. About one-third of the people with diabetes received treatment with insulin.
Hypoglycemia was defined as a blood sugar reading of less than 50 milligrams per deciliter (mg/dl). The researchers found that 7.7 percent of the hospitalized people with diabetes had at least one episode of low blood sugar.
The researchers controlled the data for a number of factors, including the presence of other illnesses, and predicted what a normal length of stay for that condition should be.
"Even after including all of that, low blood sugar still gave us additional information about the prediction of mortality that we're not capturing with all the other metrics," said Turchin.
The risk of inpatient death jumped 85.3 percent for each additional day with hypoglycemia, the team found. Up to one year later, the risk of death was still increased by 65.8 percent for each additional day someone experienced hypoglycemia. The researchers also found that the length of a hospital stay increased by 2.5 days for each day there was at least one low blood sugar reading.
"If patients develop low blood sugar, they should be closely monitored for clinical deterioration," said Turchin. "Keeping blood sugar high doesn't have a benefit, but being too aggressive in lowering it may not be such as a good idea either."
Dr. Joel Zonszein, director of the clinical diabetes center at Montefiore Medical Center in New York City, said that his sense is that the higher risk of mortality in people with hypoglycemia is a result of underlying disease, rather than from aggressive treatment.
"There's no question that hypoglycemia is associated with a high mortality rate, but it's an association; there's no direct evidence of causality," he said.
All rights reserved