Obstructive Sleep Apnea and
Glucose Metabolism
Naresh M. Punjabi, MD, PhD
Johns Hopkins University, School of Medicine
Departments of Medicine and Epidemiology
Baltimore, Maryland (USA)
General Outline
• Background
– Disease definition and epidemiology (e.g., glucose metabolism)
• Association between sleep apnea and:
– Insulin resistance and glucose intolerance
– Insulin secretion
• Effects of intermittent hypoxia and sleep fragmentation
on glucose metabolism
• Effects of CPAP treatment on glucose metabolism
Fasting glucose
Disease Definition: Metabolic Dysfunction
Diabetes and Glucose Tolerance
DIABETES
126
IFG
IFG + IGT
Normal
IGT
100
140
200
2-hour post-challenge glucose
Diabetes Prevalence: 2010
Diabetes Prevalence: 2030
Diabetes and IGT
Age-adjusted Percentage of U.S. Adults Who Were Obese
or Who Had Diagnosed Diabetes
Obesity (BMI ≥30 kg/m2)
1994
No Data
<14.0%
2008
2000
14.0-17.9%
18.0-21.9%
22.0-25.9%
>26.0%
Diabetes
1994
No Data
2008
2000
<4.5%
4.5-5.9%
6.0-7.4%
7.5-8.9%
CDC’s Division of Diabetes Translation. National Diabetes Surveillance System
available at http://www.cdc.gov/diabetes/statistics
>9.0%
General Outline
• Background
– Disease definition and epidemiology (e.g., glucose metabolism)
• Association between sleep apnea and:
– Insulin resistance and glucose intolerance
– Insulin secretion
• Effects of intermittent hypoxia and sleep fragmentation
on glucose metabolism
• Effects of CPAP treatment on glucose metabolism
The Sleep Heart Health Study:
Field Sites
Minneapolis (1085)
▲
South Dakota (201)
▲
Phoenix (201)
▲ Tucson (911)
▲
New
▲ York (760)
▲
Pittsburgh (398)
▲
Sacramento (501)
▲
Framingham (1000)
▲
Hagerstown (1184)
▲
Oklahoma (200)
National Heart Lung and Blood Institute (NHLBI). The Sleep Heart Health Study: Manuals of Operation.
http://www.jhucct.com/shhs/details/manual/demographics/01jul02received/shhs1demo1jul02.pdf. Accessed July 18, 2007.
Sleep Apnea and Fasting Glucose Values
Percentage
20
17.5
Impaired
Diabetic
15
12.1
10
8.8
8.7
7.2
5
4.0
0
< 5.0
5.0 - 14.9
RDI (events/h)
Punjabi et al. Am J Epidemiol. 2004;160:521.
15.0 +
Sleep Apnea and Glucose Tolerance
50
Percentage
40
30
Impaired
Diabetic
36.0
31.2
29.1
20
10
12.3
9.3
15.0
0
< 5.0
5.0 - 14.9
RDI (events/h)
Punjabi et al. Am J Epidemiol. 2004;160:521.
15.0 +
Sleep Apnea and Fasting Glucose
Predictor
Fasting Glucose Level
(n = 2,656)
Odds Ratio
95% CI
Respiratory disturbance index (no. of events/hour)
< 5.0
1.00
5.0 – 14.9
1.27
0.98 - 1.64
≥ 15.0
1.46
1.09 - 1.97
Average oxyhemoglobin saturation during sleep (%)
≥ 95.72
1.00
94.57 – 95.71
1.52
1.05 - 2.20
93.32 – 94.56
1.75
1.21 - 2.53
< 93.32
1.95
1.34 - 2.84
Punjabi et al. Am J Epidemiol. 2004;160:521.
Sleep Apnea and Insulin Resistance
HOMA
= Go x Io
HOMA Index (Units)
3.6
P = 0.008
P = 0.002
3.2
2.8
Is insulin resistance
enough for diabetes?
2.4
2.0
< 5.0 5.0–14.9 ≥ 15.0
RDI
Adjusted for age, gender, smoking status, BMI, waist circumference,
and sleep duration
HOMA = homeostasis
Punjabi et al. Am J Epidemiol. 2004;160:521.
model assessment
Sleep Apnea and Insulin Resistance:
HOMA and Oxygen Saturation
3.4
Quartiles
I : < 93.32%
II : 93.32% - 94.56%
III : 94.57% - 95.71%
IV : > 95.72%
HOMA Index (Units)
3.2
3.0
2.8
**
2.6
*
2.4
*P = 0.04
2.2
**P = 0.01
2.0
(for comparisons with
the first quartile)
I
II
III
IV
Quartiles of Average
Saturation During Sleep
Punjabi et al. Am J Epidemiol. 2004;160:521.
17/23
25/26
Glucose Intolerance and Diabetes:
Two Defects
Genes
Insulin Secretion
Insulin Resistance
Environment
Glucose Intolerance
Type 2 Diabetes
Acute Insulin Response to IV Glucose:
Normal Subjects
Plasma Insulin (U/mL)
100
Glucose
80
60
40
20
0
–30
0
30
Time (min)
Adapted from Robertson & Porte. J Clin Invest. 1973;52:870-876, with permission.
Acute Insulin Response to IV Glucose:
Normal and Type 2 Diabetic Subjects
Plasma Insulin (U/mL)
100
Glucose
100
80
80
60
60
40
40
20
20
0
Time –30
(min)
0
Normal
30
Glucose
0
Time –30
0
30
(min)
Type 2 Diabetes
Adapted from Robertson & Porte. J Clin Invest. 1973;52:870-876, with permission.
Changes in acute insulin response (AIR)
relative to changes in Insulin Sensitivity
500
AIR (µU/ml)
400
300
NGT
IGT
200
100
NGT
NGT
NGT
DM
0
0
1
Weyer C et al. J Clin Invest 1999;104:787–794
2
3
Insulin Sensitivity
4
5
Insulin Sensitivity and Insulin Secretion
in Sleep Apnea
• 118 subjects
• Men 71; Women 47; 86.4% White
• No medical conditions
• Mean age was 45.7 years (range: 23 – 73)
• Mean BMI was 29.4 kg/m2 (range: 17.2 – 52.2).
• DEXA percent body fat 32.1% (range: 10.1 – 60.5).
Punjabi et al. ARJCCM. 2008; In press
Insulin and Glucose Profile: IVGTT
300
Normal Subject
(AHI = 1.2/hr)
Glucose
200
400
Insulin
300
150
200
100
100
50
0
0
20
40
60
80
100
120
140
160
0
180
Insulin (U/ml)
Glucose (mg /dl)
250
500
Insulin and Glucose Profile: IVGTT
800
350
Sleep Apnea Subject
(AHI = 72/hr)
250
700
Glucose
600
Insulin
500
200
400
150
300
100
200
50
100
0
0
20
40
60
80
100
120
140
160
0
180
Insulin (U/ml)
Glucose (mg /dl)
300
Insulin Sensitivity in Sleep Apnea
6.0
SI ([mU/L]-1[min]-1)
Test for linear trend across groups: p < 0.0007
5.0
4.0
3.0
2.0
1.0
< 5.0
5.0 - 14.9
15.0 - 29.9
> 30.0
Apnea-hypopnea index (events/hr)
Insulin Sensitivity and Oxygen Desaturation
10
SI ([mU/L]-1[min]-1)
8
6
4
2
0
0
2
4
6
Average DSaO2 (%)
8
10
AIRG ([mU/L][min])
Insulin Secretion in Sleep Apnea
Apnea-hypopnea index (events/hr)
Punjabi and Beamer. ARJCCM (2009)
Integrated b-cell Function in Sleep Apnea
Disposition Index
Test for linear trend across groups: p < 0.034
Apnea-hypopnea index (events/hr)
Punjabi and Beamer. ARJCCM (2009)
Alterations in Glucose
Metabolism in Sleep Apnea
• Independent of total body fat, obstructive sleep apnea is
associated with insulin resistance, glucose intolerance,
type 2 diabetes
• Obstructive sleep apnea may also impair the
compensatory response in insulin secretion for a given
degree of insulin resistance
Punjabi and Beamer. ARJCCM (2009)
Pathogenesis of Metabolic Abnormalities
Sleep Apnea
Hypoxia
?
Glucose Intolerance
Insulin Resistance
Type 2 Diabetes
Arousals
Human Experimental Approach
• Two distinct experimental paradigms
– Effects of acute intermittent hypoxia in normal
subjects
• 5-hour exposure during wakefulness
– Effects of sleep fragmentation in normal subjects
• Two nights of sleep disruption with auditory and mechanical
stimuli (~60/hr)
Effects of Acute Intermittent Hypoxia on Glucose
Metabolism in Awake Normal Subjects
21%
5%
Louis and Punjabi. Journal of Applied Physiology (2009)
Effects of Acute Intermittent Hypoxia on Glucose
Metabolism in Awake Normal Subjects
• Study Protocol
– Hypoxia day
• 5% O2 continued until O2 saturation reaches 85%
• 21% O2 continued until O2 saturation reaches baseline level
(95-97%)
– Normoxia day
• 21% O2 delivered throughout the 8-hour period
• Manual two-way valve used to alternate from one room air
tank to another
Louis and Punjabi. Journal of Applied Physiology (2009)
Start IVGTT
(~1:30 pm)
Start Protocol
(~8:30 am)
End Protocol
(~4:30 pm)
100
90
SaO2 (%)
80
70
60
5 minutes
21
FiO2 (%)
5
100
77
SaO2 (%)
96
96
97
80
79
60
EtCO2 (%)
5
0
EKG during one episode of desaturation
Louis and Punjabi. Journal of Applied Physiology (2009)
Effects of Acute Hypoxia:
Insulin Sensitivity and Insulin Secretion
p < 0.0179
4.0
3.0
2.0
1.0
0.0
p = 0.85
500
AIRG ([mU/L][min])
SI ([mU/L]-1[min]-1)
5.0
400
300
200
100
0
Normoxia
Intermittent Hypoxia
Normoxia
Louis and Punjabi. Journal of Applied Physiology (2009)
Intermittent Hypoxia
Effects of Acute Intermittent Hypoxia on Glucose
Metabolism in Awake Normal Subjects
• Acute intermittent hypoxia for as little as 5-hours during
wakefulness
– Decreases insulin sensitivity
– Not associated with a compensatory increase in insulin secretion
– Decrease glucose effectiveness
– Increases sympathetic nervous system activity
– Not associated with any changes in serum cortisol
Louis and Punjabi. Journal of Applied Physiology (2009)
Am J Respir Crit Care Med. 2004
Increases in cortisol and catecholamines
Sleep Fragmentation in Normal Subjects
Follow-up IVGTT
Baseline IVGTT
Sleep Fragmentation
Day 1
Habituation
Night
Day 2
Fragmentation
Night
Day 3
Fragmentation
Night
N = 11
Stamatakis and Punjabi (Chest - 2010)
Day 4
Sleep Fragmentation in Normal Subjects
Sleep Fragmentation in Normal Subjects
Stamatakis and Punjabi (Chest - 2010)
Sleep Fragmentation in Normal Subjects:
Insulin Sensitivity and Insulin Secretion
5.0
4.0
3.0
2.0
1.0
p < 0.001
p = 0.08
AIRg ([mU/L][min])
SI ([mU/L]-1[min]-1)
6.0
600.0
400.0
200.0
0.0
0.0
Baseline Post-Fragmentation
Baseline Post-Fragmentation
Stamatakis and Punjabi (Chest - 2010)
Effects of Sleep Fragmentation on Glucose
Metabolism in Normal Subjects
• Sleep fragmentation (non-specific) for two nights
– Decreases insulin sensitivity (Si)
– Increases insulin secretion to compensate for lower Si
– Decrease glucose effectiveness
– Increases sympathetic nervous system activity
– Increase morning cortisol levels
Stamatakis and Punjabi (Chest - 2010)
Tasali E. et.al. PNAS 2008;105:1044-1049
Tasali E. et.al. PNAS 2008;105:1044-1049
Mechanistic Links: Sleep Apnea and
Metabolic Dysfunction
Sleep Fragmentation
Sympathetic Activation
Sleep Apnea
HPA dysregulation
Insulin
Resistance
Systemic Inflammation
b-cell
Dysfunction
?
Intermittent
Hypoxemia
Type 2 Diabetes
Effects of CPAP on Insulin Sensitivity
Insulin Sensitivity Index (ISI) assessed with the hyperinsulinemic
clamp at baseline, after 2 d, and after 3 m of CPAP therapy
ISI (Whole group, n = 40)
(After 3 months, n = 31)
(mol/kg · min)
Baseline
5.75 + 4.20
After 2 days CPAP therapy
6.79 + 4.91
Improvement compared to baseline
After 3 months CPAP therapy
Improvement compared to baseline
P = 0.003
7.54 + 4.84
P = 0.001
Harsch et al. Am J Respir Crit Care Med. 2004;169:156.
Type 2 Diabetes, Glycemic Control, and Continuous
Positive Airway Pressure in Obstructive Sleep Apnea.
Babu et al. Arch Intern Med, 2005
7/19
Sleep
Metabolic
Dysfunction
A bi-directional relation
Periodic Breathing
Cheyne-Stokes Respiration
T2DM and sleep
• Sleep is disturbed in T2DM subjects
• Obstructive sleep apnea is common in those
with T2DM but mediated primarily by obesity
• Periodic or Cheyne-Stokes breathing is common
in T2DM and may contribute to potential added
morbidity and mortality
Sleep
Metabolic
Dysfunction
A bi-directional relation
Conclusions and Summary
• Independent of obesity, sleep apnea is associated with
insulin resistance, glucose intolerance, and type 2
diabetes
• Intermittent hypoxemia and recurrent arousals may
mediate the metabolic abnormalities in sleep apnea
• CPAP perhaps mitigate the metabolic disturbance?
(more research is still needed)
• Diabetes in turn may predispose to breathing
abnormalities during sleep