Background and aims: Diabetes mellitus affects about 7.5% of the adult population.
Diabetic keto-acidosis (DKA) and the hyperglycemic hyperosmolar
state (HHS) are the most serious acute complications of diabetes. They are
frequent within the diabetic population and tend to increase with the growing
diabetes pandemic. These acute complications remain associated with a
high mortality rate (<5% for DKA and 15% for HHS) despite an optimal care.
The aim of the study was to analyze the precipitating factors, complications,
treatment, length of stay (LOS) and mortality of these hyperglycemic crises in
the university hospital of Lausanne.
Materials and methods: This study describes a retrospective analysis of 173
patients admitted for DKA, HHS or both in the department of Internal Medicine
(Emergency Room, Intensive Care Unit, Ward) between 1995 and 2006.
Demographic factors, admission diagnoses and laboratory data were examined.
The acute hyperglycemic crises were defined according to the 2006 ADA
criteria, i.e. DKA as a plasma glucose ≥13,9 mmol/L, an arterial pH ≤7,30
with the presence of either blood or urine ketones; HHS as a plasma glucose
≥33,3 mmol/L together with an effective serum osmolality ≥320mOsm/kg.
We defined mixed DKA-HHS crises as the combination of both.
Results: Median age of the study population was 50 (33,5-67,8) years, 61%
were males. At the admission, 32% of cases presented ACD, 32% HHS and 35%
mixed DKA-HHS. The main precipitating factors were non-adherence (49%),
infections (26%), inaugural diabetes (24%), idiopathic (8%) and cardiovascular
(4%). The median LOS was 9 (6-15) days, type 2 diabetic patients had a longer
LOS (14 vs. 6 days, p<0.0001). DKA patients had a shorter median hospital
LOS than other patients (7 vs. 10 days, p=0.022). In-hospital mortality was the
highest in mixed DKA-HHS (8.2%), followed by DKA (5.4%) and HHS (3.6%).
Older age was a predictive factor for mortality (77 vs. 53 years, p=0.011), as well
as in multivariate analysis (p<0.0001). Consultation by a diabetes team was associated
with much lower mortality rates (1.8% vs. 58.3%, OR=0.04, p<0.0001).
A consultation by a specialized diabetes team during the first 3 hospitalization
(early) days significantly reduced LOS in both univariate (7 vs. 15,5 days,
OR=0.35, p<0.0001) and multivariate analyses (R2=0.46, p<00001).
Conclusion: Patient's non adherence with their diabetes treatment and inaugural
diabetes are the major precipitating factor for DKA, HHS and mixed DKA-HHS
crises, which was never previously described in HHS or mixed DKA-HHS crises.
We observe a shift in the distribution of the precipitating factors, which were previously
outclassed by infections. Mortality was highest in the mixed DKA-HHS
crises group, a situation that can be accounted for by the greater metabolic unbalance
in this group. We highlight that an early consultation by a diabetes team
highly decreases the LOS. Thus, we suggest that diabetic patients' education be
reinforced, the diabetes team be involved as soon as the first hospitalization days
and patient's education program for emergency team could be useful.