Effect of diabetes and glucose variability on fall-risk in the hospital population
C. Berra, F. Zangrandi, F. De Fazio,  M. Mirani, G. Favacchio M. Albini. P. Meroni, F. Folli

Background
Patient’s falls are the most frequent adverse events in hospitals, with severe consequences such as the compromise of physical and mental well-being and increased in-patient time which further increases economic and social costs. Previous studies have suggested that severe hypoglycemia is associated with risk of falls, both in type 1 as well as type 2 diabetic subjects, irrespective of age.. Elderly subjects with and without diabetes also have increased risk for falls, fall-related fractures, seizures and comas and exacerbation of chronic conditions, such as cognitive dysfunction and cardiac events which leads to hospitalization. Little is known about how  glycemic variability and other risk factors could be associated with falls in hospitalized population. 

Aim of the study
To investigate the role of hypoglycemia (glucose<70 mg/dl) or hyperglycemia (glucose>200 mg/dl), or the combination of both, as independent risk factors for falls in a hospitalized population

Material and Methods
From January 2014 in our hospital all patients have electronic notes with all the clinical and laboratory events are recorded, including capillary glucose and falls.
We have made a retrospective analysis on all patients admitted in our hospital from January 2015 to December 2016 in every ward except the intensive care units. All capillary glucose values were analyzed (when present) in patients who had a fall and those which did not fall (fall population and non-fall population). 
We have considered the presence of value of capillary blood glucose >200 mg/dl and <70 mg/dl, as an indication of glycemic variability, in the hypoglycemic or hyperglycemic range.
The obtained data were subjected to multivariate analysis matched for sex, age, admission to hospital (from emergency care or elective) surgical or medical admission and discharge diagnosis.

Results
We analyzed medical records of 57411 patients, of which 759 had a fall during the hospital stay. The number of patients at least with one capillary glucose determination in the non fall population (N=56652) was 13025 (23%); while in the fall population (N=759) was 385 (50,7%). (p<0.001)
In the group which did not have any capillary glucose measurements (N=43627) there were 374 falls (0.8%). In the group which had at least one capillary glucose measurement (N=13410) there were 385 falls (2.9%) (p<0.001)
Among the fall-population subgroup, which had at least one capillary glucose measurement, 243 subjects out of a total of 385 (63,1%) had a value less than 70 mg/dl or higher than 200 mg/dl, while in the non fall population,  6412 out of 13025 subjects (49,2%). had values out of the 70 to 200 mg range 
Patients with at least one glycemic value out of the range (>200 mg/dl and <70 mg/dl) had a 1.76 odd ratio of falls as compared to the well compensated subjects (glucose values 70-200 mg/dl) (confidence interval between 1.42 and 2.19; p <0.001)
Considering the use of diabetes medications, in subjects treated with hypoglycemic agents, including insulin, 233 out of 7579 (3.07%) had a fall, while in the subjects not receiving hypoglycemic agents, the percentage of falls was 1.05% (526 on 49832) (odds ratio 2.97, confidence interval between 2.54 and 3.49; p <0.001).
After multivariate analysis for sex, age, admission to hospital (from ER/elective) surgical or medical and discharge diagnosis, the presence of at least one glycemic value out of the range confers an odd-ratio for fall of 1.5 (confidence interval between 1.20 and 1.86; p <0.001).

Discussion
Diabetic subjects with or without a history of severe hypoglycemia are more likely to fall than those without diabetes. Not only in elderly patients, but this risk is more pronounced in people younger than 65 y/o, and previous studies on the relation between glycemic control and the risk of falls have been inconclusive. 
Our data suggests that not only hypoglycemia but also modest hyperglycemia (with values only more than 200 mg/dl) are indicators of risk for falls among the hospitalized population. This risk is increased if the degree of metabolic compensation is worse with a direct relation between the number of value of glycemic blood test out of range and risk of falls. This increased risk is independent from for sex, age, admission to hospital (from ER/elective), surgical or medical, and severity of discharge diagnosis.

Effect of diabetes and glucose variability on fall-risk in hospital population