Table S1 Search

Database Search

Cochrane (March 20, 2014)

EMBASE (1974-March 20, 2014)

MEDLINE (1948-March 20, 2014)

PubMed (March 20, 2014)

(sleep apnea or sleep apnoea or sleep apnea, obstructive or obstructive

sleep apnea or obstructive sleep apnoea) and (exhaled condensate

biomarker* or salivary biomarker* or urinary biomarker* or blood

biomarker* or serum biomarker* or biomarker*)

LILACS (March 20, 2014) "apnea" or "apneia" and "marcadores biologicos"

Google Scholar (March 20, 2014) “biomarkers and sleep apnea” (without patents or citations)

1

Table S2

Excluded articles and reasons for exclusion (n=26).

Author, year Reasons for

exclusion

Bratel et al 1999 15 6

Calvin et al 2010 16 1

Cholidou et al 2013 17 2

Culla et al 2010 18 2

El-Solh 2002 19 2

Gozal et al 2007 20 2

Gozal et al 2010 21 2

Khalyfa et al 2011 22 3

Kim et al 2012 23 2

Kishida et al 2014 24 8

Lin et al 2013 25 5

Loubaki et al 2008 26 7

Makino et al 2006 27 7

Osorio et al 2014 28 2

Oyama et al 2012 29 2

Patel et al 2009 30 4

Przybylowski et al 2006 31 5

Roche et al 2009 32 2

Rubinsztajn et al 2006 33 5

Salord 2014 34 2

Staats et al 35 4

Uysal et al 2014 36 2

2

Van Hoorenback 2012 37 6

Wang et al 2005 38 5

Wang et al 2010 39 2

Wang et al 2012 40 4

1-Central apnea, 2-OSA with morbidities, 3- Not OSA patients, 4-Different target condition, 5- Not English,

Spanish or Portuguese, 6-Not full overnight PSG, 7-Not biomarkers study, 8- Review.

3

Table S3

Complementary information regarding included pediatric articles (n=35).

Year Author Findings Main Conclusion

2002 Gozal et al 42 Serum VEGF concentrations of >100 pg/ml were weakly predictive of OSA. Circulating VEGF levels were frequently

elevated in OSA.

2004 Tauman et al 43 CRP levels were significantly higher in the OSA group. Plasma CRP levels were increased in OSA

and correlated with AHI.

2005 Kaditis et al 61 Insulin and HOMA index values were similar in children with AHI≥5 and with

AHI<5 (P>0.05).

Severity of OSA was not a significant

predictor of fasting insulin or HOMA index

values.

2005 Larkin et al 44 CRP levels varied with increasing BMI and OSA. AHI≥5 was associated with increasing

levels of CRP.

2006 Goldbart et al 45 LTB4 and cys-LT (LTC4/LTD4/LTE4) levels were elevated in OSA when

compared to mild OSA and control. PGE2 concentrations were similar among

the three groups.

Increases in leukotriene concentrations

may be a noninvasive tool in the clinical

assessment.

2006 Kheirandish-

Gozal et al 46

Mean initial CRP levels decreased after adenotonsillectomy (P < 0.05). OSA is frequently associated with increases

in CRP levels that are reversible upon

treatment.

2006 Krishna et al 47 Gelsolin, perlecan are differentially expressed in urine of OSA children. Increased expression of gelsolin and

4

perlecan in the urinary proteome of OSA

children could potentially serve as

biomarker.

2006 Li et al 66 Moderate to severe OSA group had significantly greater fasting insulin levels

and HOMA compared to the other two groups.

OSA is prevalent in obese children with HS

and insulin is independently associated with

the condition.

2006 Montgomery-

Downs et al 48

IsoP-m values were unrelated to any polysomnographic measures. Oxidative stress is not a significant feature

of pediatric OSA.

2006 O’Brien et al 49 P-selectin levels were significantly higher in the OSA group and the mild OSA

group when compared to control.

OSA children have plasma elevations of P-

selectin. Elevations in ICAM-1 are primarily

associated with obesity rather than OSA.

2006 Shah et al 50 Proteomic patterns were capable of diagnosing OSA with 93% sensitivity and

90%specificity.

Proteomic profiling of serum samples in

OSA children revealed differential

expression of circulating proteins that may

provide useful future diagnostic

approaches.

2007 Kaditis et al 62 Overnight change in cICAM-1 was similar in OSA children compared to those

with mild OSA or to children with an index less than or equal to 1 (P>0.05).

Overnight change in cICAM-1 levels was

not related with severity of OSA.

2007 Tauman et al 51 Log leptin concentrations were significantly lower in controls (P = 0.006) and

in children with SpO2 nadir ≥90% than children with SpO2 nadir <90%, even

OSA and associated hypoxemia may

5

after controlling for BMI z score (P<0.03). No significant differences were

found in log resistin levels as a function of obesity or AHI. Significant

correlations between log adiponectin levels and log Insulin/Glucose (I/G)

ratios (-0.28, P = 0.006) and between log leptin levels and log I/G ratios (r =

0.66, P<0.0001) emerged.

contribute to the elevation of leptin levels.

2008 Gozal et al 52 IL-6 levels were higher and IL-10 plasma levels were lower in OSA. They

returned to control levels after tonsillectomy and adenoidectomy surgery.

Systemic inflammation is a consequence of

OSA, even in the obesity absence, and is

reversible upon treatment.

2008 Li et al 68 Moderate OSA (OAI>5) had significantly higher CRP levels compared to non-

OSA (P=0.01). OAI was independently associated with CRP (P=0.001).

Sixteen children underwent treatment and there was significant reduction in

their serum CRP after intervention [P=0.033]. A significant correlation was

also demonstrated between change in CRP and change in OAI following

treatment for OSA.

OSA children may have associated

systemic inflammation as reflected by a

raised CRP that decreased significantly

following treatment.

2009 Gozal et al 53 Combinatorial approaches indicated that the presence of values beyond the

calculated cutoff concentrations for 3 or more of the proteins yielded a

sensitivity of 95%and a specificity of 100%.

Pediatric OSA is associated with specific

and consistent alterations in urinary

concentrations of specific proteins.

2009 Kaditis et al 63 Moderate-to severe OSA had higher log-transformed urine Cys-LTs levels

than those with mild OSA, PS, or control (P < 0.05).

Urine excretion of Cys-LTs is related to

pediatric OSA severity.

6

2009 Kaditis et al 64 Children with severe hypoxemia had significantly higher log-transformed

norepinephrine levels compared to those with moderate hypoxemia (P<0.05)

or compared to controls (P<0.05).

Severity of nocturnal hypoxemia in children

with intermittent upper airway obstruction

during sleep correlates with morning urine

levels of norepinephrine suggesting

increased sympathetic tone.

2010 Kim et al 54 Plasma log MRP8/14 levels showed AHI dose-dependent increases

regardless of obesity. Log MRP8/14 levels correlated with log AHI (P<0.001)

after controlling for age and body mass index Z-score, and with endothelial

function.

Plasma MRP8/14 levels were associated

with OSA.

2010 Li et al 67 OSA children did not have significantly different adiponectin and leptin

concentrations than non-OSA for both the obese and non-obese groups.

Systolic BP, age, high-density lipoprotein cholesterol, and BMI z -score were

independently associated with adiponectin, whereas diastolic BP, triglyceride,

height, and BMI z -score were independently associated with leptin

concentration.

BMI rather than OSA was the main

determinant of adipokines.

2011 Bhushan et al 55 Morning plasma FABP4 levels were increased in OSA children. Of the 11

SNPs tested, the frequency of rs1054135 (A/G) minor allele (A) was

significantly increased in OSA and obese subjects.

Childhood obesity and OSA are associated

with higher plasma FABP4 levels and thus

promote cardiometabolic risk.

2012 DeBoer et al 41 Sleep study parameters hsCRP or ΔhsCRP overnight. The degree of insulin resistance correlated

7

significantly with increased overnight

increase in hsCRP, while measures of

oxygen desaturation and AHI were not

correlated.

2012 Khalyfa et al 56 Morning plasma MIF levels were increased in OSA children. Of the 28 SNPs

tested, the frequency of rs10433310 minor allele was significantly decreased

in OSA. The minor allele frequency of all other 27 SNPs was similar in OSA

and non-OSA groups.

Pediatric OSA is associated with higher

plasma MIF, hsCRP, and fasting insulin

levels that promote cardiometabolic risk,

and the MIF gene SNP rs10433310 may

account for some of the variance in such

risk.

2012 Malakasioti et al

65

Children with moderate-to-severe OSA had higher log-transformed H2O2

concentrations compared to mild OSA, or to control. No significant

differences were demonstrated between the three study groups in terms of

EBC NOx levels.

Children with moderate-to-severe OSA

have increased H2O2 levels in morning

EBC, an indirect index of altered redox

status in the respiratory tract.

2012 Stefanini et al 74 Hemoglobin, hematocrit, and HDL were all significantly higher in the Snore

group when compared to the OSA group. VLDL levels were higher in the

OSA group. There was no statistical difference between the groups based on

OSA severity.

Non-obese OSA children present no

significant alterations in metabolic tests or

BP levels.

2013 Benedek et al 75 EBC pattern of OSA was discriminated from control (P = 0.03) EBC biomarker pattern can be

8

distinguished in OSA from HS.

2013 Gozal et al 57 OSA children and abnormal endothelial function had significantly lower

adropin concentrations compared with controls (P < 0.001) and OSA children

and normal endothelial function (P < 0.001). Individual adropin concentrations

were not significantly correlated with age, BMI z-score, obstructive AHI, or

nadir oxygen saturation. Mean adropin concentration measured after

adenotonsillectomy in a subset of OSA children showed an increase in the

OSA+/ EF+ group (P < 0.01), but essentially no change in the OSA+EF-

group (P > 0.05).

Plasma adropin concentrations are reduced

in pediatric OSA when endothelial

dysfunction is present, and return to within

normal values after adenotonsillectomy.

Assessment of circulating adropin

concentrations may provide a reliable

indicator of vascular injury.

2013 Kheirandish-

Gozal et al 58

Overnight increases in epinephrine, norepinephrine, and GABA levels

emerged in children with OSA; taurine levels decreased. Using combinatorial

approaches and cutoff values for overnight changes of these four

neurotransmitters enabled prediction of OSA (p< .0001). GABA and taurine

alterations, as well as overnight reductions in phenylethylamine, were more

prominent in OSA children.

Increases in GABA levels and decreases in

taurine levels could underlie mechanisms of

neuronal excitotoxicity and dysfunction.

Combinatorial approaches using defined

cutoffs in overnight changes in

concentrations of selected

neurotransmitters in urine may predict OSA.

2013 Kim et al 59 OSA children had significantly higher TREM-1 and pentraxin-3 levels versus

controls: P< 0.01, P< 0.05, respectively).

Plasma TREM-1 and pentraxin-3 levels are

elevated in pediatric OSA.

2013 Park et al 69 m-sCor significantly decreased with OSA severity. This decrease resulted in

a r-sCor that was significantly different between the control group and the two

Salivary cortisol may be a useful biomarker

9

OSA subgroups. n-sCor did not show a significant change with OSA severity of OSA.

2014 Jeong et al 70 Postoperative m-sCor, the difference of cortisol level (sub-sCor: m-sCor

minus n-sCor), and the ratio of cortisol level (r-sCor: m-sCor/n-sCor) showed

significant difference postoperatively.

Tonsil and adenoid surgery can normalize

disturbed cortisol secretion. Assessment of

salivary cortisol in OSA children may be a

useful tool for postoperative management

planning and follow-up.

2014 Kheirandish-

Gozal 60

Non-obese controls had the lowest levels of LBP, and the presence of obesity

without OSA was associated with significant LBP increases. Non-obese

children with OSA exhibited increased LBP levels, with obese children with

OSA demonstrating the highest LBP levels of all four groups.

Systemic low-level endotoxemia and

resultant systemic inflammation is present

in children who are either obese or suffer

from OSA and is particularly prominent

when both conditions are present.

Disrupted sleep and other factors facilitating

obesity such as a high-fat diet may disrupt

the gut microbiome and lead to increased

systemic LPS levels with resultant

inflammation, promoting downstream

metabolic dysfunction.

2014 Park et al 71 The sAA subtraction and ratio (P = 0.014 and P < 0.001, respectively) were

significantly higher in severe OSAS than in the mild/moderate and control

Screening test of sAA in children suspicious

of having sleep disorderd breathing may be

10

groups. Although ODI and AHI were significantly associated with sAA, sAA in

the OSAS group was not related to lowest oxygen saturation or

adenotonsillar hypertrophy

used as a helpful tool to identify those with

OSA, when PSG cannot be available.

2014 Patacchioli et al

72

Compared with controls, both mild and moderate to severe OSA children

showed: increased salivary cortisol diurnal production, and no changes in a-

amylase diurnal trajectory and production. Morning salivary cortisol

concentrations were negatively associated with the disease severity in the

moderate to severe OSA group.

Pediatric OSA is associated with

dysregulation of the HPA axis activity.

2014 Villa et al 73 Urinary 8-isoprostane levels were significantly higher in the group with AHI of

≥5 than in the group with AHI of <5 (P<0.01).

Urinary 8-isoprostane may be used as a

specific inflammatory marker to predict

OSA severity.

* All terms that mean obstructive sleep apnea (SDB, SRDB, OSAS) were standardized as OSA.

Abbreviations

AHI=apnea/hypopnea index, BMI=body mass index, BP=blood pressure, cICAM-1=circulating intercellular adhesion molecule 1, CRP=C reactive protein, Cys-

LTs=cysteinyl leukotrienes, EBC=exhaled breath condensate, FABP4=fatty acid binding protein 4, GABA=y–aminobutyric acid, H2O2=hydrogen peroxide,

HDL=high density lipoprotein, HOMA=homeostatic model assessment, HPA=hypothalamic-pituitary-adrenal axis, HS=habitual snoring, hsCRP=high-sensitivity C-

reactive protein, ICAM-1=intercellular adhesion molecule-1, IL-10=interleukin-10, IL-6=interleukin-6, IsoP-m=isoprostane metabolites, LBP=lipopolysaccharide-

binding protein, LPS= lipopolysaccharide, LTB4=leukotriene B4, LTC4=leukotriene C4, LTD4=leukotriene D4, LTE4= leukotriene E4, m-sCor=salivary cortisol after

PSG, MIF=macrophage migration inhibitory factor, MRP8=myeloid-related protein-8, n-sCor=salivary cortisol before PSG, NOx=nitrate mono-nitrogen oxides,

11

OAI=obstructive apnea index, ODI=oxygen desaturation index, OSA=obstructive sleep apnea, PGE2=prostaglandin E2, PSG=polysomnography, r-sCor=salivary

cortisol ratio, sAA=salivary alpha amylase, SNPs=single nucleotide polymorphisms, SpO2=peripheral capillary oxygen saturation, sub-sCor=subtract salivary

cortisol, TREM-1=triggering receptor expressed on myeloid cells-1, VEGF=vascular endothelial growth factor, VLDL=very-low-density lipoprotein.

12

Table S4

Complementary information regarding adults selected studies.*

Year Author Findings Main Conclusion

2000 Chin et al 99 Soluble E-selectin level had decreased after 3 to 4 days of nasal CPAP therapy,

(P =0.002) and after 1 month (P = 0.02). After 6 months, soluble vascular cell

adhesion molecule-1 levels had not changed significantly, while the mean soluble

intercellular adhesion molecule-1 level had decreased further (P=0.02). Before

treatment, soluble intercellular adhesion molecule-1 levels and the apnea and

hypopnea index were correlated (r = 0.43, P = 0.04).

OSA and hypopnea have a significant

adverse effect on serum soluble cell

adhesion molecule-1 levels that may

be reduced by nasal CPAP treatment.

2002 Carpagnano

et al 149

Higher concentrations of IL-6 were found in OSA patients (8.7 ± 0.3 pg/mL) than

in healthy control (1.6 ± 0.1 pg/mL; p < 0.0001). Obese subjects also had higher

levels than healthy control subjects, but lower levels than OSA patients (2.1 ±0.2

pg/mL, P < 0.05 and p < 0.0001 respectively). 8-isoprostane levels were found to

be higher in OSA patients (7.4 ±0.7 pg/mL) than in obese subjects (5 ± 0.3

pg/mL; P =0.4) and healthy subjects (4.5 _±0.5 pg/mL; P < 0.005)..

The inflammation and oxidative stress

are characteristic in the airway of OSA

patients and their levels depend on the

severity of the OSA. The measurement

of IL-6 and 8-isoprostane levels may

prove to be useful in screening and

monitoring obese patients who have a

high risk of developing OSA.

2002 Gozal et al 42 Serum VEGF levels were significantly higher in adults when compared to those

with mild or no disease (p<0.0001).

Circulating VEGF levels are frequently

elevated in OSA patients, and may

13

play a role in the regulation of tissue

oxygen delivery.

2002 Lavie et al

Study 1 110

AHI was found to be a significant independent predictor of morning VEGF

concentrations.

VEGF may contribute to the long-term

adaptation of OSA to recurrent

nocturnal hypoxia.

2002 Lavie et al

Study 2 110

VEGF concentrations were found to be significantly higher in OSA patients

comparing with control 1 and control 2. (P<0.007)

VEGF may contribute to the long-term

adaptation of OSA to recurrent

nocturnal hypoxia.

2002 Lavie et al

Study 3 110

A significant decrease in VEGF concentrations was found only in patients in

whom nocturnal hypoxia improved after treatment. (P<0.01)

VEGF may contribute to the long-term

adaptation of OSA to recurrent

nocturnal hypoxia.

2002 Schulz et al

114

Group A had significantly (P <0.01) increased VEGF serum levels when

compared with Group B and Group C (mean ±SEM: 410 ±77 pg/ml versus 224

±38 pg/ml and 245 ±61 pg/ml). The degree of nocturnal oxygen desaturation in

OSA significantly correlated with the VEGF concentrations (r =0.67, P <0.01).

Serum levels of VEGF are elevated in

severely hypoxic patients with OSA

and are related to the degree of

nocturnal oxygen desaturation.

2002 Shamsuzzam

an et al 88

Plasma CRP levels were significantly higher in patients with OSA than in controls

(P<0.0003). In multivariate analysis, CRP levels were independently associated

with OSA severity (F=6.8, P=0.032).

OSA is associated with elevated levels

of CRP. The severity of OSA is

proportional to the CRP level.

2003 Carpagnano Higher concentrations of 8-isoprostane were found in the EBC (9.5 ±1.9 pg/mL) Systemic and local oxidative stress are

14

et al 155 and plasma (9.7± 1.5 pg/mL) of OSA patients compared to controls (p < 0.0001).

Significant reduction of 8-isoprostane was seen after CPAP therapy (p < 0.005).

A positive correlation was found between morning exhaled 8-isoprostane levels

and the AHI (r =0.8; P < 0.0001).

increased in OSA patients, and that

they are higher after nocturnal apnea

and reduced by CPAP therapy.

2003 Christou et al

113

The measurement of antioxidant capacity did not differ between the OSA patients

and healthy sample. Patients with severe had linearly negative correlation

between antioxidant capacity in their blood samples and AHI (R =2 0:551,

P=0:041).

Reduced antioxidant capacity in serum

is an index of excessive oxidative

stress. Patients with severe OSA have

reduced values of antioxidant capacity.

2003 Ohga et al 100 nCPAP decreased apnea, desaturation, and the circulating ICAM-1 and IL-8

levels in OSA patients. The circulating levels of ICAM-1, IL-8, and MCP-1 in

untreated OSA patients were significantly greater than those in the controls.

nCPAP therapy could reduce OSA-

induced hypoxia and generation of

inflammatory mediators.

2003 Yokoe et al 101 Levels of CRP and IL-6 were significantly higher in patients with OSA than in

controls (CRP P<0.001, IL-6 P<0.05). IL-6 production by monocytes was also

higher in patients with OSA than in controls (P<0.01).

nCPAP significantly decreased levels of both CRP (P<0.0001) and IL-6

(P<0.001) and spontaneous IL-6 production by monocytes (P<0.01).

Levels of CRP and IL-6 and

spontaneous production of IL-6 by

monocytes are elevated in patients

with OSA but are decreased by

nCPAP.

2004 Guilleminault

et al 89

Analysis of variance indicated a significant difference between the groups for

diastolic BP, RDI, lowest SaO2, and BMI. The mean serum CRP level was normal

in all 3 groups.

Obesity is a risk factor for high serum

CRP levels in OSA patients, as in the

general population.

2004 Imagawa et al No significant increase in IL-6 or TNF-α was detected in the present study cohort. The elevation in VEGF is not directly

15

102 related to IL-6 or TNF-α · levels.

2005 Alzoghaibi et

al 115

Mean SOD and lipid peroxidation concentrations of patients were not significantly

different from those of control subjects (0.29±0.015 vs 0.31±0.01 U/ml and

4.64±0.57 vs 4.62±0.54 mmol/ml, respectively). Higher concentrations of IL-8 and

GCP-2 were found in OSA patients (198.8±4.76 vs 180.83±3.38 and

383.34±46.19 vs 218±13.16 pg/ml, respectively, P<0.005).

There is a significant increase in

neutrophil chemokines IL-8 and CGP-2

in serum f OSA adults.

2005 Sukegawa et

al 103

During the CPAP treatment, both 24-h urinary adrenaline and noradrenaline were

significantly lower compared with natural sleep. CPAP significantly decreased the

AHI, DSI, % stage 1, and arousal index and significantly increased the lowest

SpO2. There were no significant differences in % stage 2, % stage 3/4, and %

REM between before and during CPAP treatment. Multiple analysis of covariance

tests revealed that lowest SpO2was the most important factor for increasing 24-h

urinary noradrenaline levels (F = 4.75, p = 0.048).

One night CPAP treatment could

improve autonomic dysfunction.

2005 Yamauchi et

al 104

Urinary 8-OHdG excretion was significantly higher in the severe OSA group (p

=0.03). Urinary 8-OHdG excretion was significantly correlated with parameters of

OSA. Only ODI was significantly correlated with urinary 8-OHdG excretion after

adjustment for confounding factors that are considered to be related to oxidative

stress.

The severity of OSA is independently

associated with oxidative stress.

Among various OSA parameters, ODI

is most closely related to oxidative

stress.

16

2006 Braga et al 116 S100B was higher in OSA (0.15 ± 0.09 mg/l) than in the control group (0.08 ±

0.06 mg/l; P<0.01). Serum NSE was similar in both groups (17.5 ± 12.2 vs. 15.8

± 6.8 ng/ml).

There is an elevated serum S100B

levels in OSA adults compared to

controls.

2006 Htoo et al 90 The degree of NF-κB activation was positively correlated with indices of apnea

severity. In five severe OSA patients, 1 month of CPAP therapy decreased

neutrophil NF-κB activation to control levels. sEselectin and sVCAM

concentrations were reduced by CPAP in four of these five subjects. OSA leads

to NF-κB activation, which may constitute an important pathway linking OSA with

systemic inflammation and cardiovascular disease.

Activation of NF-kB may play a central

role linking the pathophysiological

features of OSA with adverse

cardiovascular consequences.

2006 Lentini et al

117

The mean baseline CK level was significantly higher in patients with severe OSA

compared to those with mild-to-moderate OSA and controls. ROC analysis

identified an optimal cutoff value of > 148 U/L (r =0.660) for CK, which yielded a

PPV of 99%, a sensitivity of 43%, and a specificity of 95% for the diagnosis of

OSA. CPAP treatment resulted in a significant decline of CK levels both in

patients with mild-to-moderate OSA (P < 0.001) and in patients with severe OSA

(P < 0.001).

One third of the sample showed a

mild-to-moderate elevation in CK level,

which was highly predictive of OSA.

The application of CPAP therapy in

OSA patients resulted in a significant

decrease in CK level.

2006 Mehra et al 91 Linear regression analysis showed that after adjustment for subject

characteristics, waist circumference, and comorbidities, OSA was not significantly

associated with morning IL-6 levels. In contrast, linear regression analyses

showed that, compared with the participants without OSA, those with OSA had

Morning sIL-6R levels demonstrated

stronger associations with moderate to

severe OSA than morning IL-6 levels.

Associations with OSA and morning

17

significantly higher morning sIL-6R levels (mean±SD, 4.60±1.42 ng/mL

[P=0.001]), even after adjustment for subject characteristics, waist circumference,

and comorbidities, which persisted after adjustment of evening sIL-6R levels.

sIL-6R levels persisted even after

adjustment for waist circumference,

cardiovascular disease, and evening

sIL-6R levels, suggesting the potential

utility of sIL-6R as a marker for

measuring overnight OSA stresses.

2007 Peled et al 118 There was no association of VEGF or VEGF/platelets with the severity of OSA..

Age was the only parameter to significantly predict VEGF and VEGF/platelets on

multivariate analysis (R 2 = 0.713, P = 0.001 and R 2 = 0.844, P= 0.001,

respectively).

The elevation of serum VEGF in OSA

is not associated with the severity of

the disease, but it is associated with

patient age.

2007 Phillips et al

151

Compared with baseline on CPAP, withdrawal from therapy resulted in an

immediate return of OSA with an increase in RDI to 26.7 ±5.2 and 39.0 ± 5.9

events per hour after one and seven nights without CPAP, respectively (both P <

0.0001). This was accompanied by a concomitant rise in daytime urinary

noradrenaline (P < 0.0001) after seven nights CPAP withdrawal that was

positively associated with the severity of hypoxaemia. In contrast, withdrawal

from CPAP therapy was not accompanied by any change in measured cytokines

or VEGF (all P > 0.1).

One week of CPAP withdrawal was

associated with a return of OSA and a

marked increase in sympathetic

activity without a concomitant elevation

of vascular inflammatory markers.

2007 Punjabi et al

92

A strong association was found between degree of OSA and serum levels of

CRP, with or without adjustment for age and several measures of adiposity.

Mechanisms other than adiposity per

se could contribute to the inflammatory

18

state seen in OSA adults.

2007 Ryan et al 119 CRP levels were similar in groups 1, 2 and 3 (p = 0.727), but were significantly

higher in group 4 than in the other groups (p<0.05 by individual group

comparisons). There was no difference in homocysteine levels between all four

groups (p = 0.1). CPAP did not alter CRP (P = 0.145) or homocysteine levels (P

= 0.381).

CRP and homocysteine levels are not

associated with OSA severity in men.

2007 Ursavas et al

120

Circulating levels of both ICAM-1 (480.1 ±216.7 vs. 303.4 ± 98.6 ng/ml, p !

0.0001) and VCAM-1 (1,156.6 ±79.8 vs. 878.8 ±71.1 ng/ml, p = 0.002) were

significantly increased in the OSA group compared to the control group. For an

ICAM-1 cutoff level of 375 ng/ml, predictive sensitivity and specificity for OSA

were 69.2% (95% confidence interval, CI: 52.4–83.0%) and 82.4% (95% CI:

65.5–93.2%), respectively. For a VCAM-1 cutoff level of 859 ng/ml, predictive

sensitivity and specificity for OSA were 74.4% (95% CI: 57.9–86.9%) and 64.7%

(95% CI: 46.5–80.2%), respectively.

OSA can independently increase

circulating levels of adhesion

molecules.

2007 Ye et al 76 Serum concentrations of CRP and MMP-9 were significantly higher in OSA

patients than in controls. Levels of CRP and MMP-9 were significantly higher in

patients with moderate to severe OSA than in patients with mild OSA or in obese

control subjects. A positive correlation was found between levels of CRP and

MMP-9 in OSA patients.

AHI, mirroring the frequency of IHR,

was a predictor of enhanced circulating

CVD biomarkers MMP-9 and CRP.

IHR contributes to the upregulation of

the inflammatory factors in OSA

patients

19

2008 Antonopoulou

et al 156

An increased level of leptin and respective increase of inflammatory variables

was found. No significant association was observed between parameters of EBC

and plasma leptin levels. A part of the parameters of disease severity is

significantly associated with pH and 8- isoprostane. Smoking did not seem to be

a critical confounding factor for evaluation of the above measurements.

There is no significant association

between the increased levels of leptin

and increased airway inflammation in

OSA.

2008 Arias et al 152 Nocturnal urinary levels of norepinephrine, epinephrine and sTNFR-1 (1,053±269

versus 820±166 pg.mL-1) were significantly higher in OSA patients. There were

no significant differences in plasma levels of IL-6, LTB4, or TNF-α between the

two study groups. There were no significant differences in BP, urinary

catecholamine levels, or plasma IL-6, LTB4 and TNF-a levels after both treatment

modalities. After 3 months of effective CPAP usage, sTNFR-1 levels were

significantly reduced (1,053±269 vs 899±254 pgmL-1).

OSA patients have higher levels of

sTNFR-1 than Non-OSA individuals.

sTNFR-1 levels are lowered by CPAP

therapy.

2008 Burioka et al

105

Serum IL-6 levels were significantly reduced after CPAP therapy by 46% (p

<0.005). No significant 24 h variation of serum IL-6 in severe OSA patients was

found before CPAP; however, a significant 24 h variation of serum IL-6 was

found after CPAP.

Intermittent hypoxia during sleep may

contribute to systemic inflammation

and result in an elevation of serum IL-6

in severe OSA patients.

2008 Constantinidis

et al 121

Production of TNF-a and IL-6 were significantly elevated in OSA patients and

obese controls compared with overweighted control subjects (p < 0.05). Serum

levels of IL-1β did not differ among the study groups. Preoperative cytokine

values were significantly correlated with the preoperative BMI and the AHI in

Surgical management of mild to

moderate OSA is indicated for the

reduction of pro-inflammatory

cytokines.

20

OSA patients. Surgery resulted a significant reduction in the TNF-α and IL-6

values of the study group. Decrease in cytokine level was strongly correlated with

the AHI decrease. The postoperative relative percentage change of IL-6 values

was significantly higher than this of TNF-α (P < 0.001). Surgical management of

mild to moderate OSA leads to a significant reduction in TNF-α and IL-6 values.

2008 Kanbay et al

122

There was a significant negative correlation between plasma TNF- α and

adiponectin levels in OSA group. Compared with the non-obese OSA group,

subjects with obesity and OSA had lower adiponectin levels and SpO2 <90%, and

higher TNF- α levels. Obese OSA patients had higher rates of CVD with lower

plasma adiponectin levels when compared with obese control subjects.

Serum adiponectin is significantly

lower in OSA patients and it is

independent of obesity.

2008 Li et al 157 IL-6 and TNF-α in EBC and serum gave the highest correlation coefficients (r =

0.62 and r = 0.71 in EBC; r = 0.58 and r = 0.66 in serum, respectively) as well as

the lowest AIC values (63.87, 68.97; 62.65, 70.64, respectively).

IL-6 and TNF-α measurements may be

used in OSA treatment follow-ups,

when PSG is not available.

2008 Norman et al

93

Markers of hypoxia (lowest oxygen saturation level and %T<90), correlated

significantly with AST and ALT levels (Pearson’s r = -0.31 to -0.38, P <0.003),

while AHI, BMI, BP, fasting glucose, triglyceride, and cholesterol levels did not.

Hierarchical linear regression was then done to determine the best predictors of

aminotransferase levels. Markers of metabolic syndrome were entered as one

block and markers of sleep apnea as another. Regression analyses explained

16.3% of the variance in AST and 18.9% of the variance in ALT, with %T<90

In OSA patients serum ami-

notransferase levels are better

predicted by markers of oxygen de-

saturation than by factors traditionally

associated with the metabolic

syndrome.

21

playing the largest role.

2008 Petrosyan et

al 150

The levels of eNO and eCO were higher in OSA patients than in control subjects

(P<0.05). Nasal NO was higher in OSA patients than in obese controls (P<0.01).

The level of H2O2, 8-isoprostane, LTB4, and nitrates were elevated in OSA

patients in comparison with obese subjects (P<0.01). pH was lower in OSA

patients than in non-apneic controls (P<0.01). One month of CPAP therapy

increased pH (P<0.05) and reduced eNO (P<0.001) and nNO (P<0.05). AHI was

positively correlated with 8-isoprostane (r=0.42; P<0.05), LTB4 (r=0.35; P<0.05),

nitrates (r=0.54; p<0.001), and H2O2 (r=0.42; P<0.05).

Airway inflammation and oxidative

stress are present in the airway of

OSA patients in contrast to non-apneic

subjects. Exhaled breath markers are

positively correlated with the severity

of OSA. One-month administration of

CPAP improved airway inflammation

and oxidative stress.

2008 Takahashi et

al 106

The TRX level was significantly higher (p =0.02) and the adiponectin level was

significantly lower (P= 0.02) in the OSA group than in the non-OSA group. After 1

month of nCPAP (n =27), the TRX level significantly decreased (P =0.03), and

the adiponectin level significantly increased (p =0.03). Among the 14 patients

with untreated OSA, the TRX and adiponectin levels did not significantly change

over a 1-month interval. Among the 53 (41 OSA =12 non-OSA) subjects, the TRX

level was positively correlated with the respiratory disturbance index (p =0.001)

and percentage of time with SaO2 <90% (p =0.0002). The adiponectin level, but

not the TRX level, was correlated with the BMI (n =53; P =0.02).

Plasma TRX may be a unique marker

for evaluating oxidative stress and

monitoring the effectiveness of nCPAP

in OSA patients.

2008 Zamarron et

al 123

OSA patients presented significantly higher circulating levels of PAI-1 compared

with the control group, and the difference was even more marked in patients with

OSA patients presented higher

circulating levels of PAI than controls

22

both OSA and hypertension. OSA patients presented a significant inverse

correlation between PAI-1 levels and AHI (r = -0.71, P<0.001).

group, which was even greater when

patients had associated hypertension.

2009 Kim et al 124 Multiple regression analysis showed that haptoglobin and apolipoprotein M levels

are independently related to AHI (P < 0.01).

Of these nine proteins, haptoglobin,

paraoxonase and apoliprotein M were

quantified in 74 subjects.

2009 Kuramoto et

al 107

Serum level of SAA (P<0.05), brachial-ankle PWV (P<0.05) and BP (P<0.005)

were significantly higher in Group 3 than in Group 1.

Markers of inflammation and

autonomic dysfunction are increased in

patients with OSA, and nCPAP might

help to reduce risk factors for

cardiovascular disease.

2009 Lam et al 124 Urinary catecholamines were positively correlated with severity of sleep apnea,

independent of obesity. BP measurements correlated with age, obesity, severity

of sleep apnea, and urinary catecholamines. Regression analysis showed that

sleep indices and urinary catecholamines were independent determinants of

morning systolic and diastolic BP, respectively, while total cholesterol and waist

circumference were respective additional factors. Urinary catecholamines and

waist circumference were determinants of evening BP, with morning cortisol

being an additional determinant for diastolic BP.

OSA and related sympathetic activity

contributed significantly to the

determination of daytime BP in

overweight middle-aged men without

overt cardiometabolic diseases, and

other contributing factors include

abdominal obesity, total cholesterol,

and cortisol levels.

2009 Lederer et al Median plasma KL-6 levels were higher in patients with OSA compared with

controls: 317 (232–506) UmL-1 versus 226 (179–257) UmL-1, respectively. Higher

Circulating KL-6 levels are elevated in

some patients with OSA, possibly

23

94 plasma KL-6 levels were associated with greater time spent asleep with an

oxyhaemoglobin saturation <90%, lower nadir saturation, more frequent

desaturation of >4% during sleep and lower brachial artery flow-mediated

dilation. Adjustment for nadir saturation or flow-mediated dilation attenuated the

association between plasma KL-6 levels and OSA.

reflecting increased alveolar wall

permeability.

2009 Li et al 78 Biomarker levels, in both EBC and serum, differed significantly across the four

groups. Classification by LDA using IL-10 in EBC showed the highest agreement

with AHI classification (kappa = 0.88). Logistic regression distinguished moderate

and severe OSA from mild OSA and non-OSA perfectly using IL-6 in EBC and

almost perfectly using IL-10 in EBC (area under the ROC curve = 0.997). The

levels of biomarkers among smokers overlapped with mild to severe OSA

patients.

EBC IL-6 and IL-10 have potential to

predict severity of OSA in non-smoker

OSA suspects.

2009 Lui et al 79 After adjustment for age, smoking, BMI, waist circumference, and sleep

efficiency, CRP correlated positively with the AHI [r =0.35, P < 0.001], duration of

O2 saturation < 90% (r =0.29, P =0.002), and arousal index (r =0.32, P =0.001),

and it correlated negatively with minimal O2 saturation (r = -0.29, P = 0.002).

These correlations were consistent when adjustment was made for magnetic

resonance image visceral fat volume instead of waist circumference. In the

regression model, significant predictors of CRP included AHI, waist

circumference, and triglycerides (adjusted R2, 0.33, p =0.001, p =0.002, P

In healthy middle-aged men, elevated

CRP level is associated with OSA

independent of visceral obesity.

24

=0.018, respectively). Among the 111 subjects, 32 subjects with no or mild OSA

(AHI < 15 events/h) were matched with 32 subjects with moderate-to-severe OSA

(AHI > 15 events/h) in magnetic resonance image visceral fat volume. CRP was

higher in subjects with moderate-to-severe OSA (median, 1.32; 0.45 to 2.34

mg/L) when compared to subjects with no or mild OSA (median, 0.54; 0.25 to

0.89 mg/L; P = 0.001).

2009 Ting et al 111 BMI, neck circumference, waistline circumference, respiratory disturbance index,

arousal index, lowest oxygen saturation, duration of SaO2 < 90%, blood glucose,

hs-CRP, and metabolic syndrome score in Morning Surge and Constant High

were significantly greater than those in Constant Low. Except metabolic

syndrome score, all other parameters in Morning Drop were similar to those in

Constant Low.

OSA patients whose post- to pre-

overnight SSBPs were elevated or

maintained a constant high have more

sleep respiratory disturbance, more

pro-inflammatory state, and higher

metabolic syndrome indices than the

rest. Without subdividing into Constant

Low, Morning Drop, Constant High,

and Morning Surge, the important

pathophysiologic points of OSA

patients will possibly be missed.

2009 Ucar et al 125 Morning lactate was significantly higher in the OSA group than the Non-OSA

group (1.65 ±0.48 and 1.35 ±0.57 mmol/L, respectively) (P =0.003). Lactate

levels at night and the change overnight in lactate levels were not significantly

As a marker of tissue hypoxia, arterial

lactate may be used to assess the

severity of OSA.

25

different between groups. After an adjustment for age, gender, and BMI, lactate

levels before PSG were related to the AHI (beta: 0.004, 95% CI: 0.000–0.008)

and the rate of sleep-time spent under 90% oxygen saturation (T90%). The

following morning lactate level was correlated with the T90% (beta: 0.005, 95%

CI: 0.000–0.010). After an adjustment for lactate levels before PSG, lactate in the

morning was correlated with T90% (beta: 0.004, 95% CI: 0.000–0.008).

2010 Lee et al 126 Twenty-three percent of the overall patients had significantly elevated serum

levels of hs-CRP. The increase of hs-CRP correlated fairly with BMI, AHI, tonsil

size, and ESS (r =0.450, 0.398, 0.393, and 0.300, respectively; all P ≤0.05) after

adjustment for conventional coronary heart disease risk factors. Only the AHI

could predict for significantly elevated hs-CRP after stepwise multiple linear

regression (R2 =0.251, P <0.001).

Patients in this study with hs-CRP≥ 3

mg/L were more prevalent in the

severe OSA group. This observation

suggests that the severe OSA patients

need to have their cardiovascular

statuses evaluated by use of an hs-

CRP screening test.

2010 Pallayova et

al 95

Both increases in AHI and the presence of prediabetes/diabetes were associated

with reductions in HOMA-IS in the entire cohort even after adjustment for sex,

race, age, and BMI (P = 0.003). In subjects with NGM (n = 30), OSA severity was

associated with significantly increased HOMA-B (a trend towards decreased

HOMA-IS) independent of sex and adiposity. OSA-related oxyhemoglobin

desaturations correlated with TNF-α (r=-0.76; P = 0.001) in women with NGM

and with IL-6 (rho=-0.55; P = 0.035) in women with IGM (n = 15) matched

OSA is independently associated with

altered glucose homeostasis and

increased basal beta cell function in

severely obese adults with NGM.

Moderate to severe OSA imposes an

excessive functional demand on

pancreatic beta-cells, which may lead

26

individually for age, adiposity, and AHI. to their exhaustion and impaired

secretory capacity over time. The two

distinct biomarker profiles linking sleep

apnea with NGM and IGM via TNF-a

and IL-6 have been discerned in our

study to suggest that sleep apnea and

particularly nocturnal oxyhemoglobin

desaturations are associated with

chronic metabolic fluxes and specific

cytokine stressors that reflect links

between sleep apnea and glucose

metabolism.

2010 Steiropoulos

et al 127

OSA patients had significantly higher TNF-α levels (P < .001) while no other

difference in the examined inflammation markers was recorded between groups.

Overall, TNF-α levels were correlated with neck circumference (P < .001), AHI (P

= .002), and Oxygen Desaturation Index (P = .002).

Obese OSA patients have elevated

TNF-α levels compared to BMI

matched controls, suggesting a role of

OSA in promoting inflammation,

possibly mediated by TNF- α.

2010 Ye et al 80 Serum DNA, IL-6, and MDA concentrations were measured and were

significantly higher in patients with moderate and severe OSA groups than those

in the mild OSA and control groups (p< 0.05). Univariate analysis showed that

The increasing concentration of serum

DNA in patients with OSA was

positively correlated with disease

27

serum DNA correlated positively with AHI, ODI, IL-6, and MDA, and negatively

correlated with minimal oxygen saturation (miniSaO2) (all p<0.05). In stepwise

multiple regression analysis, only MDA and miniSaO2 were suggested as

significant independent predictors for the serum DNA concentrations. After 6

months of nCPAP therapy, serum concentrations of DNA, IL-6, and MDA were

significantly decreased (P<0.05).

severity. Serum DNA may become an

important parameter for monitoring the

severity of OSA and effectiveness of

therapy.

2011 Akinnusi et al

96

pLOX-1 levels were higher in subjects with OSA than in control (326.9 ±267.1

pg/mL and 141.1 ± 138.6 g/mL, respectively; P = 0.004).

Patients with OSA showed a threefold increase in baseline endothelial

expression of LOX-1 relative to control subjects.

CPAP therapy resulted in a significant decrease in endothelial LOX-1 expression

only in CPAP-adherent patients.

Circulating apoptotic endothelial cells correlated directly with baseline expression

of LOX-1 (R 2 =0.32, P = 0.01) after adjustment for age, BMI, and waist to hip

ratio.

The increased burden of circulating

apoptotic endothelial cells in OSA

adults is associated with upregulation

of LOX-1 expression on vascular

endothelium. Adherence to CPAP

therapy ameliorates these

derangements and pay reverse the

long-term cardiovascular complications

of OSA.

2011 Cintra et al 128 The cysteine plasma levels were higher in patients with OSA compared with

control (490.16 ± 67.00 m mol/L vs 439.81 ± 76.12 m mol/L, respectively, P

<0.01). The homocysteine plasma levels did not differ between groups. Cysteine

plasma levels were higher in the OSA lean subgroup when compared with lean

control (484.21 ± 71.99 μmol/L vs 412.01 ±70.73 μmol/L, respectively, P =0.009).

Cysteine is a potential biomarker of

OSA and is reduced after effective

OSA treatment.

28

There was a significant decrease of cysteine plasma levels after 6 months of

CPAP effective therapy.

2011 Jurado-

Gamez et al

130

The OSA patients showed a significant worsening of morning IRH, and a

significant increase in malondialdehyde and 8-hydroxydeoxyguanosine levels.

Only the oxygen desaturation index independently explained morning IRH, while

malondialdehyde levels showed a weak effect on IRH. In severe OSA patients,

IRH improved significantly after CPAP treatment, as did malondialdehyde, 8-

hydroxydeoxyguanosine and protein carbonyl levels.

In OSA patients, endothelial

dysfunction and oxidative stress were

observed, and IRH worsened after

sleep. The increase in oxidative stress

was not associated with IRH, while

intermittent hypoxia was strongly

associated with IRH. In severe OSA

patients, CPAP treatment improved

oxidative stress and endothelial

function.

2011 Kohler et al

153

CPAP withdrawal led to a recurrence of OSA within a few days and a return of

subjective sleepiness, however was not associated with significant deterioration

of psychomotor performance within 2 weeks. Endothelial function decreased

significantly in the CPAP withdrawal group compared with therapeutic CPAP

(mean difference in change, –3.2%; 95% CI, –4.5, –1.9%; P <0.001). Compared

with continuing CPAP, 2 weeks of CPAP withdrawal was associated with a

significant increase in morning systolic BP (mean difference in change, +8.5 mm

Hg; 95% CI,+1.7,+15.3mmHg; P=0.016), morning diastolic BP (mean difference

CPAP withdrawal usually leads to a

rapid recurrence of OSA, a return of

subjective sleepiness, and is

associated with impaired endothelial

function, increased urinary

catecholamines, BP, and heart rate.

29

in change, +6.9 mm Hg; 95% CI, +1.9, +11.9 mm Hg; P =0.008), and morning

heart rate (mean difference in change, +6.3 bpm, 95% CI,10.4,+12.2 bpm; P=0.

0.035). CPAP withdrawal was associated with an increase in urinary

catecholamines but did not lead to an increase in markers of systemic

inflammation, insulin resistance, or blood lipids.

2011 Ladesich et al

97

After controlling for age, sex, race, smoking, BMI, alcohol intake, fish intake, and

omega-3 supplementation, RBC DHA was inversely related with OSA severity.

For each 1-SD increase in DHA levels, a patient was about 50% less likely to be

classified with severe OSA. The OR (95% CI) was 0.47 (0.28 to 0.80) and 0.55

(0.31 to 0.99) for being in the severe group versus the none/mild or moderate

groups, respectively.

Disordered membrane fatty acid

patterns may play a causal role in OSA

and that the assessment of RBC DHA

levels might help in the diagnosis of

OSA. The effects of DHA

supplementation on OSA should be

explored.

2011 Pallayova et

al 98

Surgical weight loss resulted in significant decreases in BMI (16.7±5.97

kg/m2/median 365 days), AHI, CRP, IL-6, sTNFαR1, sTNFαR2, and leptin levels,

while ghrelin, adiponectin, and soluble leptin receptor concentrations increased

significantly. Utilizing an AHI cutoff of 15 events/h, significantly elevated levels of

baseline sTNFαR2 and greater post-WL sTNFαR2 decreases were found in

subjects with baseline AHI ≥15 events/h compared to those with AHI <15

events/h despite no significant differences in baseline BMI, age, and BMI. In a

multivariable linear regression model adjusting for sex, age, impaired glucose

Of all the biomarkers, the decrease in

sTNFα R2 was independently

determined by the amelioration of

sleep apnea achieved by bariatric

surgery. The results suggest that

sTNFαR2 may be a specific sleep

apnea biomarker across a wide range

of body weight.

30

metabolism, BMI, and follow-up period, the post- weight loss decreases in AHI

were an independent predictor of the decreases in sTNFαR2 and altogether

accounted for 46% of the variance of sTNFαR2 (P=0.011) in the entire cohort.

2011 Zamarron et

al 132

OSA patients presented higher circulating levels of ICAM-1, endothelin-1 and

PAI-1 than the control group. No differences were found in E-selectin and vWF.

After 1 year of CPAP treatment, there was a significant decrease in circulating

levels of ICAM-1 and PAI-1. No differences were found in endothelin-1, E-

selectin and vWF.

OSA is associated with elevated levels

of ICAM-1 and PAI-1 and these levels

normalize after treatment with CPAP.

2012 Duru et al 133 Median serum S100B protein level was 133.7 pg/mL (range 20.97-230.70 pg/mL)

in OSA patients and 16.1 pg/mL (range 10.1-22.9 pg/mL) in control group

(p<0.005). Serum 100B protein level did not correlate with any studied variable

(p>0.05 for each correlation coefficient).

Serum S100B protein level is

increased in OSA patients and may be

a useful maker for OSA.

2012 Feng et al 81 Serum chemerin levels were significantly elevated in OSA patients (120.93 ±

25.84 μg/L vs. 107.51 ±20.41 μg/L). Multivariable logistic regression analysis

revealed that serum chemerin levels were an independent determinant of the

presence of OSA (OR 1.030, 95% CI 1.016–1.045; P < 0.001). Serum chemerin

levels in severe OSA patients were significantly higher compared with those in

mild and moderate OSA patients (P= 0.015 and P=0.020, respectively).

Spearman correlation analysis indicated that serum chemerin levels were

correlated with the severity of OSA (r = 0.210, P=0.016). Serum chemerin were

Elevated serum chemerin levels could

be an independent predicting marker

for OSA.

31

positively correlated with waist circumference (r = 0.164, P=0.008), BMI (r =

0.158, P=0.014), systolic BP (r = 0.135, P=0.037), homeostasis model

assessment of insulin resistance (r = 0.140, p = 0.031), CRP (r = 0.202,

P=0.002), and AHI (r = 0.152, P=0.022).

2012 Guven et al

134

The serum hs-CRP levels were significantly higher in the OSA group (4.03±3.58

mg/L) than in the control group (2.41±1.95 mg/L) (P=0.013). This high level was

positively correlated with BMI (r=0.376, P=0.001) and with AHI (r=0.280,

P=0.014). In multiple regression analysis, elevated hs-CRP levels were

associated with AHI (F=3.293, P=0.033), which was independent of obesity.

OSA severity is responsible for the

elevation of hs-CRP. OSA patients

have elevated serum levels of hs-CRP,

a marker for inflammation and an

independent risk predictor for

cardiovascular morbidity.

2012 Hira et al 129 Both before and after sleep uric acid levels of patients with OSA were found to be

significantly higher (P=0.001 and 0.002, respectively) as compared to UA levels

of controls. A statistically significant (P=0.02) overnight (after sleep) rise was

observed in the serum lactate level of OSA patients. The correlation between

serum uric acid values and %TSTs below 95% SaO2 (P=0.02) was statistically

significant. The correlation was positive with %TSTs below 90% SaO2, whereas it

was found negative with normal basal oxygen. No significant correlation was

observed between serum uric acid and the AHI. Polysomnographic variables

failed to show significant correlation with serum UA on respective multiple

regression models controlling for age, BMI, and waist-hip ratio.

The measurement of serum lactate

level was a better marker of oxidative

stress among OSA patients.

32

2012 Jurado-

Gamez et al

135

The severe desaturation group showed significantly higher values in the AHI,

MDA, ICAM-1 and P-selectin (p<0.005), as well as a worsening of IRH (p=0.001).

Only ICAM-1 (p=0.019) and P-selectin (p=0.033) were independently associated

with IRH in a multiple-linear regression model.

Patients with OSA and greater

intermittent hypoxia showed worse

endothelial function, and higher levels

of MDA, ICAM-1 and P-selectin.

Nevertheless, ICAM-1 and P-selectin

rather than MDA were independently

associated with IRH.

2012 Lee et al 136 There was no significant difference in either oxidative stress or antioxidant status

markers among the three groups. There was no significant correlation between

the oxidative stress markers and the OSA variables. There were correlations

between WHR–oxLDL (r=0.424), WHR–GPX (r=0.318), WHR–TAS (r=−0.317),

and WHR–SOD (r=−0.338). In multiple regression analysis, WHR was a

significant independent variable of oxLDL, GPX, TAS, and SOD.

The oxidative stress in OSA was

related to central obesity rather than

intermittent hypoxia or respiratory

disturbances.

2012 Mancuso et al

137

OSA patients showed increased protein oxidative damage and impaired

antioxidant defenses. Patients with more severe OSA had a lower total

antioxidant capability. Preliminary data on a subgroup of patients (n = 7) treated

with CPAP show a significant increment of the FRAP values

(P < 0.005).

Oxidative stress markers may be

useful to detect and monitor redox

imbalance in OSA. FRAP might be a

new useful biomarker to monitor in vivo

the oxidative response to CPAP

therapy.

2012 Papaioannou The group median dim light melatonin onset did not differ in OSA patients The circadian phase is the same in

33

et al 112 compared to healthy subjects (OSA patients: 90 [60–150] min; healthy subjects:

135 [90–150] min, p = 0.19).

OSA patients and healthy subjects.

2012 Simiakakis et

al 138

The levels of d-ROMS were significantly higher (p = 0.005) in the control group

but the levels of antioxidant capacity were significantly lower (p = 0.004) in OSA

patients. The most important factors predicting the variance of oxidative stress

were obesity, smoking habit, and sex. Parameters of sleep apnea severity were

not associated with oxidative stress. Minimal oxygen desaturation and smoking

habit were the most important predicting factors of BAP levels.

Obesity, smoking, and sex are the

most important determinants of

oxidative stress in OSA subjects.

2012 Sokucu et al

139

RDW increased significantly with increased severity of OSA (P = 0.046) and was

positively correlated with the AHI (P = 0.002, r = 0.300), even in the non-anemic

group (P = 0.013, r = 0.291). The AHI was significantly higher in the group with

high RDW (> 15; P = 0.046). RDW was negatively correlated with sleep time (p =

0.028, r = 0.217), average oxygen saturation of hemoglobin (P = 0.003, r = -

0.239), and minimum desaturation value (P = 0.016, r = -0.235).

In patients with a clinical diagnosis of

OSA, RDW may be a marker for the

severity of the condition.

2012 Svensson et

al 131

The levels of CRP, IL-6 and lysozyme were significantly higher in subjects with

AHI ≥15 compared with women with lower AHI. All inflammatory markers except

MPO correlated to AHI and oxygen desaturation measures, and to waist

For women from the general

population there is an independent

correlation between OSA and

34

circumference. In multiple linear regressions adjusting for age, waist

circumference and smoking, independent correlations between ODI and

inflammation were found for IL-6 (P=0.03 for % sleep time with saturation <90%)

and TNF-α (P=0.03 for ODI 3%). No significant correlations were found between

AHI and inflammation.

inflammation, even after adjusting for

obesity. The results indicate that

intermittent hypoxia, and not the AHI,

is related to systemic inflammation

seen in OSA.

2013 Aihara et al

108

Each measured serum biomarker (leptin, IL-6, IL-8, TNF-α, and VEGF) was

significantly correlated with waist circumference or fat area determined by

computed tomography.

Regarding airway inflammation, sputum IL-6, IL-8, TNF-α, and VEGF were

significantly correlated with OSA severity as indicated by the respiratory

disturbance index or oxygen desaturation indices. Sputum IL-6, IL-8, TNF-α, and

VEGF were significantly related to sputum neutrophil number, and sputum IL-8

and TNF-α were related to proximal airway resistance independently of BMI.

There were no significant associations between the same biomarkers in serum

and induced sputum.

Systemic and airway inflammation in

OSA might be differently regulated by

OSA itself and comorbid obesity,

depending on the type of cytokine.

2013 Chung et al

140

In the derivation cohort, with a STOP-Bang score ≥3, the specificity for all OSA,

moderate/ severe OSA, and severe OSA was 37.0%, 30.4%, and 27.7%,

respectively. HCO3 - level of 28 mmol/L was selected as a cutoff for analysis. With

the addition of HCO3 2 level ≥28 mmol/L to the STOP-Bang score ≥3, the

specificity for all OSA, moderate/severe OSA, and severe OSA improved to

Serum HCO3 level increases the

specificity of STOP-Bang screening in

predicting moderate/severe OSA. The

authors proposed a two-step screening

process. The first using STOP-Bang

35

85.2%, 81.7%, and 79.7%, respectively. Similar improvement was observed in

the validation cohort.

score, and the second using HCO3

level.

2013 Cofta et al 141 There was a progressive increase in the concentrations of all three selectins with

the severity of OSA.

The level of plasma adhesion

molecules may be indicative of OSA

and may contribute to cardiovascular

sequelae.

2013 Ferrarini et al

142

Putative identification of 14 statistically significant features was obtained and

changes that can be related to the episodes of hypoxia/reoxygenation

(inflammation) have been highlighted.

The patterns of variation of platelet

activating factor and

lysophospholipids, together with some

compounds related to differential

activity of the gut microflora (bile

pigments and pipecolic acid) open new

lines of research that will benefit our

understanding of the alterations,

offering new possibilities or adequate

monitoring of the stage of the disease.

2013 Guo et al 82 The plasma TRX level in severe group was increased (8.62±2.14, 13.33±5.60,

14.71±5.53, and 16.10±7.34 ng/ml; P<0.05). The TRX positively related to AHI

Plasma TRX level is associated with

OSA severity and may be used as a

36

(r=0.313; p<0.05), while negatively related to the lowest O2 saturation (r=0.266;

P<0.037). The OSA patients associated with hypertension showed elevated TRX

level (17.70±6.98 vs. 13.43±5.83 ng/ml; t=2.434, P<0.018). The cutoff value of

TRX for identifying OSA was 9.39 ng/ml (sensitivity 91 %, specificity 78 %), and

its cutoff value for differentiating moderate–severe OSA from mild OSA was

11.79 ng/ml (sensitivity 75 %, specificity 65 %).

severity indicator of OSA.

2013 Hirotsu et al

143

Uric acid levels were correlated with most important risk factors for OSA, such as

AHI, desaturation time and index, SpO2, BP, cholesterol, BMI, triglycerides and

arousal, and with OSA itself. Uric acid was increased in OSA volunteers even

after controlling for all confounders. Hyperuricemic volunteers presented lower

mean and minimum SpO2 and increased desaturation index. Minimum SpO2 was

a significant predictor of uric acid levels, which in turn was considered an

independent predictor for OSA in the binary logistic model. ROC curve analysis

for establishing cut-off points for uric acid levels as a biomarker of OSA revealed

moderate sensitivity and specificity.

A strong association was found

between uric acid levels and OSA.

2013 Kurt et al 144 Hemoglobin, platelet, CRP, MPV, and RDW values did not differ between AHI

groups. PDW was significantly higher in group D (mean value 14.4±1.8) than in

group A (13.2±0.5) (P<0.001). When the four groups were compared, group D

had the lowest minimum SpO2 value [group A (89.4±3.0), B (86.7±4.2), C

(81.2±6.4), and D (68.2±13.0)]. There was a statistically significant correlation

The severity of OSA was not

correlated with CRP, MPV, and RDW.

PDW might be related markers of OSA

severity.

37

between AHI and age (r=0.35, P<0.001), BMI (r=0.31, P=0.003), PDW (r=0.28,

P=0.006), and Epworth sleepiness scale (r=0.29, P=0.007). AHI was not

correlated with CRP, MPV, and RDW. PDW is higher in severe OSA and is

correlated with different parameters of breathing function during sleep.

2013 Murase et al

109

The Ngal level correlated significantly with OSA severity as determined by the

AHI (r = 0.24, p = 0.01) and 4% ODI (r = 0.26, p = 0.01). Multiple regression

analysis showed that the Ngal level was associated with 4%ODI independently of

other clinical variables. Although the OSA (4% ODI: 33.1±16.7 to 1.1±1.9/h,

P<0.01) had significantly improved in those with good compliance, the Ngal

levels were not significantly changed (60.5±18.1 before CPAP vs 64.2±13.9

ng/ml after CPAP, P=0.27).

Plasma Ngal levels were positively

associated with the severity of OSA. It

does not seem reasonable to use the

Ngal level as a specific biomarker of

OSA in clinical practice.

2013 Ntalapascha

et al 145

The overnight change (%) of GSH/GSSG ratio and GSH was significantly

different between OSA and controls (p=0.03 and p=0.048, respectively). Plasma

protein carbonyls, erythrocyte catalase activity, 8- isoprostane, SOD, TBARS,

and TAC plasma values were not different between OSA and controls (p>0.05).

No significant correlation was found between changes in the levels of biomarkers

and AHI, arousal, or desaturation index.

OSA per se might be associated with

increased oxidative burden possibly

via GSH/GSSG pathway.

2013 Ozben et al

146

Patients with OSA had a higher incidence of hypertension and BMI but lower

serum copeptin level (P=0.007) compared with the healthy controls. There was

no significant difference regarding to serum copeptin levels between the

Copeptin levels in OSA patients could

be valuable to demonstrate impairment

in ADH regulation.

38

moderate and severe OSA patients (P=0.409).

2013 Pinto et al 154 At baseline, NOx levels showed a significant decrease during the night in both

groups (P<0.001). U-NE level and BP were significantly higher in the severe OSA

group. After 1 month of CPAP, there was a significant increase in NOx levels and

a reduction in U-NE level and BP only in patients with severe OSA.

One month of CPAP results in

significant improvements in NOx

levels, 24-h U-NE level and BP in

patients with severe OSA, but not in

patients with mild-moderate OSA.

2013 Shi et al 83 Serum S100A12 levels were significantly higher in the OSA group than in the

control group (132.17 (range 101.86 to 174.49) ng/ml vs. 78.40 (range 58.35 to

129.44) ng/ml, P<0.01). Multivariate logistic regression demonstrated that

S100A12 was the only significant and independent predictor of OSA (OD 1.012,

95 % CI 1.006 to 1.017; P<0.01). Serum S100A12 levels elevated with the

increase in the severity of OSA (S100A12 levels of 106.04 (range 83.92 to

135.13) ng/ml in mild OSA group, 133.51 (range 109.64 to 208.95) ng/ml in

moderate OSA group, and 173.04 (range 131.88 to 275.77) ng/ml in severe OSA

group; P<0.001). Serum S100A12 levels were independently correlated with AHI

scores (r=0.324, P<0.001)

Serum S100A12 levels were

independently associated with the

presence and severity of OSA. Serum

S100A12 level could be a potential

biomarker for OSA.

2013 Wang et al 85 Plasma fractalkine levels were significantly higher in patients with OSA than in

controls (463.15±110.78 versus 364.67±64.81 pg/mL, F=2.58, P=0.004).

Fractalkine were associated with AHI (r=0.756, P<0.0001), lowest oxygen

OSA is associated with elevated levels

of fractalkine. The severity of OSA is

proportional to the fractalkine level.

39

saturation (r=−0.466, P=0.005), and mean oxygen saturation (r=−0.344,

P=0.043). Plasma fractalkine levels were significantly decreased in patients with

OSA after four nights nCPAP (463.15±110.78 versus 416.75± 97.67 pg/mL,

P=0.001).

2013 Wang et al 84 Serum omentin-1 levels were significantly decreased in OSA patients compared

with healthy controls. Multivariable logistic regression analysis revealed that

serum omentin-1 levels were inversely associated with the presence of OSA

(OR= 0.520, 95% confidence interval 0.433 to 0.623; P<0.001). Severe OSA

patients had significantly lower serum omentin-1 levels compared with mild and

moderate OSA patients.

Decreased serum omentin-1 levels

could be considered as an marker of

OSA.

2013 Zhang et al 86 Compared with the control group, patients with severe OSA presented

significantly higher levels of hsCRP (1.10±0.28 vs. 0.88±0.20 mg/l) and serum

cystatin C (0.87±0.12 vs. 0.74±0.10 mg/l) (p<0.05 for all comparisons). After

adjustment for confounding factors, AHI was significantly and positively

associated with serum cystatin C levels (β00.284, P=0.007).

Serum cystatin C was associated with

the severity of OSA in younger men.

2014 Tual-Chalot et

al 147

VEGF content carried by circulating microparticles from OSA patients was

increased when compared with microparticles from non-OSA patients. Circulating

microparticles from OSA patients induced an increase of angiogenesis that was

abolished in the presence of the antagonist of endothelin-1 receptor type B.

Endothelin-1 secretion was increased in human endothelial cells treated by OSA

Circulating microparticles from OSA

patients can modify the secretome of

endothelial cells leading to

angiogenesis.

40

microparticles.

2014 Vavougios et

al 148

Statistically significant correlations were detected between DJ-1’s levels and AHI

(P=0.04), Desaturation Index (P=0.012), and LDL (P=0.042).

DJ-1 may be a useful biomarker in

OSA.

2014 Wang et al 87 Serum YKL-40 concentrations were significantly elevated in OSA patients than in

controls. Multivariate logistic regression including all variables revealed that YKL-

40 was the significant and independent predictor for the present of OSA. There is

a significant positive correlation between increments in serum YKL-40

concentrations and severity of OSA. Serum YKL-40 concentrations were

independently and significantly correlated with AHI scores.

YKL-40 could be used as a potential

biomarker for predicting OSA.

* All terms that mean obstructive sleep apnea (SDB, SRDB, OSAS) were standardized as OSA.

Abbreviations

8-OHdG=8-hydroxy-2’-deoxyguanosine, ADH=antidiuretic hormone, AHI=apnea/hypopnea index, AIC=akaike information criterion, ALT=alanine aminotransferase,

AST=aspartate aminotransferase, BMI=body mass index, BP=blood pressure, CI=confidence interval, CK=creatine phosphokinase, CPAP=continuous positive

airway pressure, CRP=C-reactive protein, CVD=cardiovascular disease, DHA=omega-3 fatty acid docosahexaenoic acid, d-ROMs= derivatives of reactive oxygen

metabolites, DJ-1=gene that is involved in tumorigenesis and in maintaining mitochondrial homeostasis, DSI=desaturation index, EBC= exhaled breath

condensate, eCO=exhaled carbon monoxide, eNO=exhaled nitric oxide, ESS=epworth sleepiness scale, FRAP=ferric reducing antioxidant power, GCP-

2=granulocyte chemotactic peptide-2, GPX=glutathione peroxidase, GSH=reduced glutathione, GSSG=oxidized glutathione, H2O2= hydrogen peroxide, HCO 3

=bicarbonate, HOMA-B= Homeostasis Model Assessment- β-cell function, HOMA-IS=homeostasis model assessment estimates of insulin sensitivity, hs-CRP=high

41

sensitivity C-reactive protein, ICAM-1=intercellular adhesion molecule-1, IGM=immunoglobulin M, IHR=intermittent hypoxia/reoxygenation, IL-6=interleukin-6, IL-

8=interleukin-8, IL-10=interleukin-10, IQR=interquartile range, IRH=ischemic reactive hyperemia, KL-6=Krebs von den Lungen-6, LDA=linear discriminant analysis,

LDL=low-density lipoprotein, LOX-1=oxidized low-density lipoprotein receptor-1, LTB4=leukotriene B4, MCP-1=monocyte chemoattractant protein-1,

MDA=malondialdehyde, MMP-9=matrix metallopeptidase-9, MPO=myeloperoxidase, MPV=mean platelet volume, nCPAP=nasal continuous positive airway

pressure, NF-kB=proinflamatory transcription nuclear factor kappa B, Ngal=neutrophil gelatinase-associated lipocalin, NGM=normal glucose metabolism,

nNO=nasal nitric oxide, NOx=reduced plasma nitrate mono-nitrogen oxides, NSE=neuron-specific enolase, ODI=oxygen desaturation index, OR=odds ratio,

OSA=obstructive sleep apnea, oxLDL=oxidized low-density lipoprotein, PAI-1=plasminogen activator inhibitor-1, PDW=platelet distribution width, pLOX-1=plasma

LOX-1, PPVs=positive predictive values, PSG=polysomnography, PWV=pulse wave velocity, RBC=red blood cell, RDI=respiratory disturbance index, RDW=red

cell distribution width, ROC=receiver operating characteristic, SAA= serum amyloid, SaO2=oxygen saturation, S100B=S100 Calcium Binding Protein B, sIL-

6R=soluble interleukin-6 receptor, SOD=superoxide dysmutase, SpO2=peripheral capillary oxygen saturation, SSBP=sleep systolic blood pressure, sTNFR-

1=soluble tumor necrosis factor receptor-1, svCAM=soluble vascular cell adhesion molecule, TAC=total antioxidant capacity, TAS=total antioxidant status, TBARS

=thiobarbituric acid-reactive substances, TNF-α R2=tumor necrose factor alpha receptor 2, TNF-α=tumor necrose factor alpha, TRX=thioredoxin, TSH=thyroid

stimulating hormone, TSTs=percentage of total sleep time spent, U-NE=urinary norepinephrine, VCAM=vascular cell adhesion molecule, VCAM-1=vascular cell

adhesion molecule-1, VEGF=vascular endothelial growth factor, vWF=von Willebrand factor, WHR=waist-to-hip ratio, WL=weight loss, YKL-40=human cartilage

glycoprotein-39.

42