Transcript STATA - Instrument Variable and Probit

The association between eGFR
and Hyperosmolar Non-Ketotic
Hyperglycemia Syndrome in
Diabetes type 2
By
Miss Jitjira Chaiyarit
ID: 567110015-4
Doctoral Seminar 1 / 2013
Outline
•
•
•
•
Introduction
Materials and Methods
Results
Discussion
Doctoral Seminar 1 / 2013
Introduction
Hyperosmolar Non-Ketotic Hyperglycemia
Syndrome (HNHS) is a life-threatening complication
of uncontrolled diabetes mellitus and often found in
patients with type 2 diabetes, is a serious condition
most frequently seen in older persons.
Complications associated with HHNS include
severe metabolic disturbances, renal impairment
etc. The testing of kidney working is an
uncommon complication of HNHS but probably
subclinical and thus underestimated.
Doctoral Seminar 1 / 2013
Materials and Methods

The design was Cross-sectional analytical.
 The information for analysis was part of
DM&HT dataset.
 The sampling method was stratified cluster
sampling with probability proportional to size
hospitals

The exclusion criteria were:
i) under age of 18
ii) not been examined serum creatinine
iii) were diagnosed with HT or both DM & HT
Doctoral Seminar 1 / 2013
Materials and Methods (Cont.)

The main outcome was HNHS (yes, no)

The factor of interest was eGFR that was
recalculated by CKD-EPI formulas
Stages of kidney disease

eGFR
Stage 1-2
≥ 60
Stage 3
30-59
Stage 4-5
≤ 29
Covariate were gender, age, receiving ACEI
or ARB and HbA1C
Doctoral Seminar 1 / 2013
Materials and Methods (Cont.)

Statistical analysis:
i) The continuous variables were presented
as mean and standard deviation.
ii) The categorical variables were presented
as frequency and percentage.
iii) Logistic regression models were used to
examine association and OR.
Doctoral Seminar 1 / 2013
Materials and Methods (Cont.)
• Ethical Consideration
The permission to study was granted by each
of sampled hospitals. Obtaining written informed
consent of all participating patients was done
prior to access to their medical records. Data
collection was done by participating hospital’s
authorized skilled personnel who had been
trained to protect and deliberately engaged in the
study.
Doctoral Seminar 1 / 2013
Materials and Methods (Cont.)
• Research frame
Doctoral Seminar 1 / 2013
Results
Baseline Characteristics
• eGFR (n=24,133)
3.0%
14.4%
Stage 1-2
Stage 3
82.6%
Stage 4-5
Mean(SD) = 84.22 (24.44)
(mL/minute/1.73m2)
Doctoral Seminar 1 / 2013
Results (Cont.)
Baseline Characteristics (Cont.)
• Gender (n= 24,263)
28.83%
• Age (n= 24,245)
37.12%
62.88%
71.17%
< 60 years
Male
Female
≥ 60 years
Mean(SD) = 56.09 (10.90)
(year)
Doctoral Seminar 1 / 2013
Results (Cont.)
Baseline Characteristics (Cont.)
• Receiving ACEI or ARB
(n= 19,162)
• HbA1C (n= 17,653)
28.93%
21.89%
71.07%
78.11%
No
Yes
< 7%
≥ 7%
Mean(SD) = 8.33 (2.14)
Doctoral Seminar 1 / 2013
Results (Cont.)
Factors associated with HNHS
Doctoral Seminar 1 / 2013
Results (Cont.)
Factors associated with HNHS adjusted for
all other factors
Doctoral Seminar 1 / 2013
Discussion
• The study found statistically significant
association between eGFR and HNHS.
• The present result found the patients with low
eGFR had risk effect (adj.OR=2.89) for having
HNHS. This is consistent with a previous finding
that if renal function declined, it was cause of
glucosuria that was a risk factor for HNHS.
Doctoral Seminar 1 / 2013
Discussion (Cont.)
• From result, age was protective effect
(adj.OR=0.80) that contrasts with research in the
past which found the average age of patients
with HNHS is 60 years. In the other hand, the
previous research has demonstrated that
HNHS may also occur in younger people.
Doctoral Seminar 1 / 2013
Discussion (Cont.)
• The patients who had poor glycemic control
(HbA1c ≥ 7%) were at greater risk (adj.OR=2.01).
This finding suggests that in those who have
not been able to control their diabetes, HNHS
will occur. This result supports past studies
show that the patients were poor management
if blood sugar levels rise and left untreated,
this can result in diabetic coma and death.
Doctoral Seminar 1 / 2013
Thank you
for your attention.