Thorarinn Gislason, Department of Respiratory Medicine and Sleep (E7), Landspitali University Hospital, 105 Reykjavík, Iceland. Tel.: +354 543 6382; fax: +354 543 6568; e-mail: firstname.lastname@example.org
The aim of this study was to investigate sleep-related sweating as a symptom of obstructive sleep apnoea (OSA). Fifteen otherwise healthy male non-smoking patients with untreated moderate-to-severe OSA underwent polysomnography, including measurements of skin and core body temperature and electrodermal activity (EDA) as an objective indicator of sweating. Evening and morning blood pressure was measured as well as catecholamines in nocturnal urine. All measurements were repeated after 3 months on successful continuous positive airway pressure (CPAP) treatment. The untreated OSA subjects had a mean (±SD) apnoea–hypopnoea index of 45.3 ± 3.9 and a mean EDA index during sleep of 131.9 ± 22.4 events per hour. Patients with higher EDA indices had higher systolic blood pressure in the evening and morning (P = 0.001 and 0.006) and lower rapid eye movement (REM) sleep percentage (P = 0.003). The EDA index decreased significantly to 78.5 ± 17.7 in the patients on CPAP treatment (P = 0.04). The decrease correlated with lower evening systolic and diastolic blood pressure (P = 0.05 and 0.006) and an increase in REM% (P = 0.02). No relationship was observed between EDA and skin or core body temperature, or to catecholamine levels in urine. OSA patients who experience sleep-related sweating may have increased blood pressure and decreased REM sleep compared with other OSA patients. CPAP treatment appears to lower blood pressure and increase REM sleep to a higher extent in these patients compared with other OSA patients.
Intense sleep-related sweating is one of the well-known clinical symptoms of obstructive sleep apnoea (OSA) experienced frequently by OSA patients (Guilleminault and Bassiri, 2005). The prevalence and pathophysiology of this symptom in adults with OSA has, however, been studied in a very limited fashion. A subjective decrease in nocturnal sweating has been found with continuous positive airway pressure (CPAP) treatment (Kiely et al., 1999). A recent study by Mold et al. (2008) found no association between subjective reports of nocturnal sweating and the apnoea–hypopnoea index (AHI) in subjects referred for a sleep study. However, an association with snoring and daytime sleepiness was reported in the same study.
Sweating is controlled solely by the skin’s sympathetic nervous system. Its main function is to increase heat loss and maintain thermoregulation. The lowering of body temperature, mainly through increased heat conduction caused by vasodilation, affects sleep onset latency to a great extent (Gilbert et al., 2004; Krauchi, 2007). This vasodilation can be assessed indirectly with the distal-to-proximal skin temperature gradient (Krauchi et al., 2000). Furthermore, body temperature has a great effect on sleep architecture, as a slight decrease (0.2 °C) in core body temperature increases deep sleep significantly (Sewitch et al., 1986) and a slight increase (0.3 °C) causes more awakenings and sleep stage changes (Fletcher et al., 1999). Thermoregulation differs between sleep stages and is less effective during rapid eye movement (REM) sleep than non-REM (NREM) sleep (Bach et al., 2002). In healthy individuals the frequency of sweating is lower in REM sleep than NREM sleep and highest in slow wave sleep (Kobayashi et al., 2003; Liguori et al., 2000).
The main focus on sleep-related sweating has concerned hot flashes during and after menopause in women. The effect of hot flashes on their sleep are still controversial, as some objective studies have shown that women with hot flashes have more arousals and awakenings, decreased sleep efficiency and REM sleep than women without hot flashes (Freedman and Roehrs, 2006; Woodward and Freedman, 1994) but other studies have shown no effect of hot flashes on sleep (Freedman and Roehrs, 2004). Hot flashes have also been associated with increases in systolic blood pressure and heart rate (Gerber et al., 2007; James et al., 2004). These studies indicate a possible role of sleep-related sweating in decreased sleep quality and increased cardiovascular risk. Hot flashes (at least a few times a week) have been reported subjectively by 4.3% of a normal male population, 55 years and older, but the timing of the hot flashes and the possible association with sleep quality and sleep disorders was not addressed (Spetz et al., 2003).
Sweating can be assessed objectively by measuring electrodermal activity (EDA), a measurement of the activation of the eccrine sweat glands (Bouscein, 1993). EDA correlates highly with an actual sweat measurement, the ventilated capsule method (r = 0.88) (Kobayashi et al., 2003).
The aim of this pilot study was to quantify sleep-related sweating in OSA subjects when untreated and again on CPAP treatment as well as to study its relationship to sleep stages, skin and body temperature and degree of respiratory disturbances. We also wanted to test the hypothesis that sleep-related sweating is related to higher blood pressure in OSA subjects.
Materials and Methods
Participants, all men who had been diagnosed with moderate-to-severe OSA (AHI = 15) were recruited from the CPAP waiting list at the Sleep Research Unit, Landspitali University Hospital, Reykjavik, Iceland. The participants were asked to undergo a new full-night respiratory sleep recording when untreated and again on successful CPAP treatment.
The exclusion criteria for the participants included current use of medication for a heart condition or hypertension, known diabetes or other chronic disorders, prior CPAP treatment and current smoking. A total of 24 OSA patients agreed to take part in the study. Nine participants were excluded from the data analyses because of CPAP non-compliance and referral to other treatment options (n = 7), predominant central apnoeas (n = 1) and current use of heart medication (n = 1). A total of 15 participants were included in the study.
The patients were on average middle-aged (49.4 ± 2.6 years), obese (body mass index 34.6 ± 2.1 kg m−2 and had a mean AHI of 45.3 ± 3.9 that decreased to 4.5 ± 0.9 with CPAP (P <0.001, Table 1). CPAP compliance on average was acceptable (5.5 ± 0.4 h) as read from the device itself. The majority of subjects used Resmed S7TM Lightweight CPAP machines (ResMed Inc., San Diego, CA, USA) and the mean group pressure was 10.3 ± 1.9 cmH2O. The consent of the National Bioethics Committee and the Data Protection Authority of Iceland was granted for this study.
Table 1. Electrodermal activity, sleep, respiratory, blood pressure and sympathetic data for untreated and continuous positive airway pressure (CPAP)-treated obstructive sleep apnoea (OSA) patients
Data are shown as mean ± SEM.
NS, not significant.
Electrodermal activity index
131.9 ± 22.4
78.5 ± 17.7
Electrodermal activity %
16.2 ± 3.3
8.3 ± 2.0
Sleep stage 1 %
9.4 ± 1.2
7.0 ± 1.0
Sleep stage 2 %
57.6 ± 1.8
52.2 ± 2.0
Slow wave sleep stage %
10.8 ± 1.0
16.9 ± 1.4
Rapid eye movement sleep %
22.2 ± 1.2
23.9 ± 1.3
25.1 ± 2.5
10.5 ± 1.3
45.3 ± 3.9
4.5 ± 0.9
Oxygen desaturation index
34.8 ± 5.0
2.1 ± 0.6
Time spent with SaO2 < 90%
53.8 ± 18.8
0.6 ± 0.2
Respiratory mechanics instability index
26.2 ± 2.1
12.7 ± 2.4
Systolic blood pressure evening (mmHg)
134.7 ± 4.8
134.8 ± 4.2
Systolic blood pressure morning (mmHg)
128.1 ± 4.4
124.5 ± 4.0
Diastolic blood pressure evening (mmHg)
80.5 ± 2.7
82.2 ± 2.3
Diastolic blood pressure morning (mmHg)
81.4 ± 2.7
77.5 ± 2.2
Adrenaline–creatinine ratio (nmol mmol−1)
3.5 ± 0.7
1.8 ± 0.4
Noradrenaline–creatinine ratio (nmol mmol−1)
21.8 ± 2.7
22.0 ± 4.5
The participants arrived at the sleep laboratory at 18:00 h after fasting (water drinking allowed) and refraining from caffeine for 4 h. Blood pressure was measured using a manual sphygmomanometer, after resting supine for 10 min. Participants had a light dinner at 19:00 h and fasted again until the study was finished in the morning. Participants were in bed from 22:00 h and awakened at 08:00 h the next morning. Nocturnal urine was collected. Blood pressure was measured again upon awakening. The same protocol was repeated after 107 ± 19 days of successful CPAP treatment.
Polysomnography (PSG) was recorded with a digital sleep recording device (EmblaTM; Flaga Inc., Reykjavik, Iceland) including two channels for electroencephalography, electrooculography, submental electromyography, electrocardiography and bilateral anterior tibialis electromyography. Nasal airflow was recorded through a cannula (nasal pressure transducer). Chest and abdominal movements were measured by respiratory inductive plethysmography belts. Pulse and oxygen desaturation were measured by a finger probe oximeter based on a four-beat exponential average (Flex Sensor 8000J and XPOD oximeter; Flaga Inc.) and body position and activity by a sensor placed on the chest.
Respiratory movements were analysed automatically to measure respiratory mechanics instability. A respiratory mechanics instability state was scored when there was a thoraco-abdominal asynchrony, as indicated by more than a 10° phase angle deviation of the thorax and abdomen (Somnologica Science 3.3.1; Flaga Inc.). A respiratory mechanics instability index was calculated as the percentage of the night spent in the state of respiratory mechanics instability.
Scoring of all recordings was performed by the same person (ESA) and used in all analyses. A qualified PSG scorer at another sleep centre in the United States made a second blind scoring of the recordings. Concordance was assessed by a two-way mixed intraclass correlation coefficient (ICC) for absolute agreement between the two scorers and calculated with a 95% confidence interval for selected variables. The concordance for respiratory disturbances was very high (ICC for the AHI 0.98 and the oxygen desaturation index 0.98). The concordance for sleep scoring was similarly high in most aspects (ICC for total sleep time 0.98, for the arousal index 0.95 and for scored minutes of REM sleep 0.93 and stages 1 and 2 sleep 0.94) but lower for slow wave sleep (ICC 0.71). Less slow wave sleep was scored by the second PSG scorer because of differences in amplitude criteria for slow wave sleep in the two sleep centres.
Measurement of EDA
Electrodermal activity was measured with NoiseFreeTM single bio-potential silver-silver/chloride electrodes (Rochester Electro-Medical, Tampa, FL, USA) connected to the EmblaTM digital recording device. EDA was recorded using the skin potential method by placing one electrode on the hypothenar eminence as an active site and another on a lightly abraded site on the volar surface of the forearm as an inactive site (two-thirds of the distance from wrist to elbow) on the right arm (Andreassi, 1995; Fowles et al., 1981). This method is believed to be a better choice physiologically than the skin conductance method, as the skin is not influenced by an external current and the recording is not disturbed by variations in the contact area (reviewed by Bouscein, 1993).
An EDA index (events per hour sleep) was calculated, considering an EDA event a change in skin potential of >50 μV amplitude and >1.5 s duration (low-cut filter 0.3 Hz) (Lajos, 2002). EDA percentage was calculated as time spent in EDA events as a percentage of total sleep time.
Measurement of temperature
Core body temperature was measured (self-inserted by the patient) 10 cm into the rectum (YSI4491E; YSI Inc., Yellow Springs, Ohio, USA) and skin temperature (YSI409B; YSI Inc.) at eight different locations on the body for the overnight recording. The YSI probes were connected to the EmblaTM digital recording device. Both core and skin temperature were measured continuously at 1-s intervals with ±0.1 °C accuracy.
Krauchi et al.’s (2000) method was used for location of skin temperature probes and calculation of the distal-to-proximal skin temperature gradient, the difference between the average temperature of distal and proximal areas. The room temperature was kept constant around 22 °C and the patients had light bed covers.
Core body temperature measurements were missing for two patients, one of whom declined use of the core body temperature probe and another whose probe was disconnected during one study.
Measurement of sympathetic urinary metabolites
Urine was collected from 22:00 h until the first post-awakening morning void (c. 10 h). As the urine collections were not 24-h samples, all values were corrected for creatinine clearance (Barratt and Topham, 2007).
Urinary adrenaline and noradrenaline were measured by high-performance liquid chromatography and adrenaline–creatinine and noradrenaline–creatinine ratios calculated. After addition of an internal standard (3,4-dihydroxybenzylamine hydrobromide), catecholamines were extracted by alumina adsorption using reagents from Chromsystems (Munich, Germany). Isocratic separations were performed on a C18 reverse phase column using electrochemical detection (Moyer et al., 1979). Measurements of urinary metabolites were missing for one patient who declined to participate in this part of the study.
The sleep period from sleep onset until waking was divided into 30-min periods and the average temperature of the first 2 h calculated to examine the lowering in core body temperature after sleep onset. The minimum 30-min core body temperature period of all the 30-min periods for each recording was also calculated. The distal–proximal skin temperature gradient, EDA index, EDA percentage, breathing parameters and arousals for each of these 30-min periods were calculated. Only periods with >20 min of sleep were used in the calculations. The Signal Workshop software was used for the analysis (SignalOne, Reykjavik, Iceland).
For the comparison of data from the untreated versus the CPAP-treated conditions, a paired sample t-test was used for parametric data while the Wilcoxon-signed rank test was used for nonparametric data. A one-way anova for repeated-measures was employed for within-subjects comparison of more than two conditions. The Pearson’s correlation was used for parametric data and the Spearman’s correlation for nonparametric data. P-values = 0.05 were considered significant.
Sleep-related sweating measured by EDA
The average EDA index (events per hour’s sleep) lowered significantly from 131.9 ± 22.4 in untreated patients to 78.5 ± 17.7 with CPAP treatment (P = 0.04, Table 1). The EDA percentage (proportional time spent in sweating) lowered from 16.2 ± 3.3 to 8.3 ± 2.0 with CPAP (P = 0.04). A highly significant correlation was found between EDA index and EDA percentage (r = 0.96 and 0.97, respectively, for untreated and CPAP-treated subjects, P <0.001). The variability in the EDA index both before and after treatment was considerable and the EDA index increased in four of 15 patients with CPAP treatment (Fig. 1). Data concerning changes in EDA, sleep and respiratory parameters, sympathetic function and blood pressure are given in Table 1.
Adrenaline levels decreased significantly in the OSA patients when they were treated with CPAP (P = 0.02), while a significant change was not observed in the noradrenaline levels. A significant decrease was found in diastolic morning blood pressure with CPAP treatment (P = 0.008) but no change in the evening or in systolic blood pressure.
When untreated, one patient out of 15 reported sweating every night or almost every night, two patients sweating three to five times a week, six patients once to twice a week, two less than once a week and four never or very seldom. The reported subjective nocturnal sweating of OSA patients lowered significantly with CPAP treatment (P = 0.008), with two reporting sweating once to twice a week, six sweating less than once a week and seven never or very seldom.
A high correlation was found between subjective measurement of sweating in untreated patients and the objective measurement of sweating as measured by the EDA index (r = 0.61, P = 0.01). This correlation was not seen for the patients on CPAP treatment.
To determine whether there was a significant difference in EDA index between sleep stages (sleep stages 1 and 2, slow wave sleep and REM), a one-way repeated-measures anova statistical analysis was conducted. Results of separate analyses for the patients when untreated and CPAP-treated showed that sleep stages influenced the EDA index (P <0.001 in both analyses) significantly. The highest EDA index was found in slow wave sleep and the lowest in REM sleep (Fig. 2).
The EDA index was lowered significantly with CPAP in all sleep stages except stage 1. The mean EDA index decreased from 146.4 to 83.9 in stage 2 sleep (P = 0.03), from 223.9 to 134.0 in slow wave sleep (P = 0.05) and from 54.0 to 20.5 in REM sleep (P = 0.04).
The amount of REM sleep as a percentage of total sleep time (REM%) in untreated patients was correlated highly with the EDA index. The higher the EDA index in untreated patients, the lower the REM% (Table 2). This relationship was maintained when controlled for the AHI. A significant correlation was also found between the EDA index in untreated patients and the change in REM% with CPAP treatment. The higher the EDA index in untreated patients, the more REM% increased (Fig. 3) and the percentage of stage 1 sleep decreased with CPAP treatment (Table 2). This relationship was maintained when controlled for the apnoea–hypopnoea index in untreated patients and the change in the index with treatment. No correlation was found between the EDA index and other sleep stages or the number of arousals and awakenings in the patients when untreated or CPAP-treated.
Table 2. Significant correlations between electrodermal activity (EDA) indices and other parameters in the obstructive sleep apnoea (OSA) patients
EDA index untreated
EDA index CPAP-treated
Δ EDA index
Significant correlations at the 0.05 level only are shown (two-tailed). Correlations are Pearson’s (parametric) or Spearman’s (non-parametric) when stated value with alphabetical letter a.
Δ: delta, change from untreated to continuous positive airway pressure (CPAP)-treated.
Rapid eye movement sleep %: untreated
Δ Rapid eye movement sleep %
Δ Sleep stage 1 %
Respiratory mechanics instability index: CPAP
Δ Respiratory mechanics instability index
Systolic blood pressure evening: untreated
Systolic blood pressure morning: untreated
Δ Systolic blood pressure evening
Δ Diastolic blood pressure evening
A relationship was found between the REM% and blood pressure. REM% in untreated subjects correlated with their systolic blood pressure in the evening (r = −0.78, P = 0.001) and morning (r = −0.65, P = 0.009). Therefore, the lower the REM%, the higher the blood pressure in untreated subjects.
EDA versus respiration, blood pressure and other sympathetic activity markers
No correlation was found between the EDA index and conventional respiratory markers (apnoea–hypopnoea index, oxygen desaturation index or the time spent in hypoxia) in the patients when untreated or CPAP-treated. However, a decrease in respiratory mechanics instability index with CPAP treatment was shown to correlate with a decrease in EDA index. On CPAP treatment, a higher respiratory mechanics instability index also correlated with a higher EDA index (Table 2).
A higher EDA index in untreated patients correlated with higher systolic blood pressure in the evening and morning in untreated patients and more decrease in evening systolic blood pressure with CPAP (Table 2). A decrease in EDA index with treatment also correlated with a decrease in evening systolic and diastolic blood pressure (Table 2 and Fig. 4).
No significant relationship was found between EDA and the nocturnal urinary adrenaline–creatinine and noradrenaline–creatinine ratios in untreated or CPAP-treated patients.
No significant difference in temperature measurements was observed between subjects when untreated and CPAP-treated. The temperature measurements included the core body temperature for the first 2 h of sleep (divided into 30-min periods), the minimum core body temperature period, the change in core body temperature from earlier time-period and the distal–proximal skin temperature gradient.
The time from the sleep onset period to the minimum core body temperature period was measured, as well as the difference in temperature (°C) between the periods for the subjects when untreated and CPAP-treated. No significant differences were found in the average time between sleep onset and the minimum core body temperature period or in the lowering of core body temperature between these two time-points.
No significant correlation was found between temperature measurements and sweating in the first 2 h of sleep. Also, no relationship was found between temperature measurements and the EDA index to breathing disturbances or arousals in the patients when untreated and CPAP-treated in the first 2 h of sleep.
The present study demonstrates that OSA patients who experience sleep-related sweating have a lower REM sleep percentage and a higher systolic blood pressure than other OSA patients. They also show a greater increase in REM sleep and a greater decrease in blood pressure with CPAP. No relationship was found between sweating and body temperature.
Sleep-related sweating in untreated and CPAP-treated patients
EDA and sleep
This study is the first, to our knowledge, to perform objective measurements of sweating in OSA patients when untreated and again on successful CPAP treatment. A major finding was the significant reduction seen in the EDA index with CPAP treatment. This reduction was not limited to patients who complained of sleep-related sweating, but was seen for most patients, independent of their sweating status. These results are in agreement with the study of Kiely et al. (1999), who found that sweating in OSA subjects decreases subjectively with CPAP treatment. Four of 15 patients did show increased sweating with treatment and none of these had a very high level of sweating when untreated. A point of interest was the great variability seen in objectively measured sweating among OSA patients both when untreated and on CPAP.
The sweating pattern in different sleep stages in the study participants is in agreement with earlier studies with healthy young participants, i.e. the least sweating was seen in REM sleep and the most in slow wave sleep (Kobayashi et al., 2003; Liguori et al., 2000). Therefore, OSA does not appear to affect the normal pattern of sweating seen in different sleep stages.
When untreated, the OSA patients appeared to be able to assess their sweating quite accurately. However, this was not the case when they were CPAP-treated. A probable explanation for this difference is that when the patients were untreated they slept more lightly and woke up more often than when on CPAP, allowing them to assess their sweating more accurately.
One of the interesting observations made in this study is the negative association found between REM sleep and sweating in OSA patients. The data that support this are both the negative correlation found between REM% and electodermal activity index in untreated patients (Table 2) and also the fact that not all the OSA patients had reduced REM% when untreated (primarily those with a high EDA index) and on average there was no change in REM% in the patients (Table 1).
Rapid eye movement sleep is possibly inhibited by a high sympathetic activity, as reflected in the high amount of sweating in untreated OSA patients. In support of this hypothesis, a negative relationship is also found between systolic blood pressure and REM%. Sympathetic activity has been found to have a role in REM sleep regulation, as both noradrenaline agonists and blockage of reuptake have been shown to suppress REM sleep in animal models (Pal and Mallick, 2006; Ross et al., 1995). A negative relationship between REM sleep and hot flashes in women has been found, strengthening these findings (Freedman and Roehrs, 2006; Woodward and Freedman, 1994).
The results of this study, therefore, show that sleep-related sweating has an additional effect on the sleep architecture of OSA patients to breathing disturbances. The patients who sweated the most had the lowest REM% when untreated and the greatest increase in REM% with treatment.
EDA versus respiration, blood pressure and other sympathetic activity markers
An association was found between the change in sleep-related sweating and the respiratory mechanics instability index with CPAP treatment, but no relationship was found between sweating and traditional markers of OSA, such as the apnoea–hypopnoea index and the oxygen desaturation index. These results are in agreement with Mold et al. (2008), in that we did not find a relationship between sweating and the apnoea–hypopnoea index. However, OSA severity possibly has an effect on sweating even though this was not seen with traditional OSA markers.
A strong correlation was found between sweating and systolic blood pressure in untreated patients. The patients who sweated the most had the highest systolic blood pressure and showed the greatest decrease in systolic blood pressure with treatment. The change in sweating with treatment was also associated with a simultaneous change in diastolic blood pressure. Sweating and blood pressure, even though controlled by different branches of the sympathetic nervous system, therefore showed similar changes with treatment. Thus, OSA subjects with a higher amount of sweating appear to have increased sympathetic activity.
No correlation, however, was found between sweating and sympathetic urinary metabolites as measured in the nocturnal sample. A 24-h collection of urine or measurements of adrenaline and noradrenaline levels in plasma could help to clarify this issue. Nevertheless, an increased sympathetic activity is not excluded, as the complexity of the sympathetic nervous system is such that changes in one aspect of the system do not necessarily reflect changes in other parts of the system.
No differences were found on average in the temperature parameters measured between untreated patients and CPAP-treated patients. The ideal way to study the temperature of OSA patients would be a 24-h constant routine measurement, as the complete circadian core body temperature rhythm would give more information regarding any subtle temperature changes and masking effects would be reduced. Sweating in OSA patients was not associated temporally with thermoregulation, neither to core body temperature nor to the distal–proximal skin temperature gradient.
Study limitations and future studies
This study was limited by its small sample size, and the results need to be ascertained in a larger cohort with a control group matched for age and body mass index. The possibility that participation in the study (due to stress or other factors) caused the observed changes therefore cannot be ruled out. However, the high correlation between subjective sweating and the objective measurement of EDA found in untreated subjects (r = 0.61, P = 0.01) and the intra-individual differences in EDA index in response to treatment (four of 15 subjects showed an increase in EDA index with CPAP) support that the changes found are indeed real. Another limitation was the possible first-night effect on measurements, as an adaptation night was not included in the study. Also, blood CO2 levels were not measured in the PSG, and hypoventilation along with the OSA cannot be excluded in the subjects. We believe this is unlikely, however, as the study group was in general good health apart from OSA. We do not believe that the different amplitude criteria for slow wave sleep used by the two scorers influenced the main findings, as data from the same scorer were used for all analyses and the concordance for other variables was high.
A study is needed that includes OSA patients with hypertension, diabetes and other comorbidities to examine whether there are differences in sweating and temperature in those subjects compared with otherwise healthy OSA patients. This type of study could prove to be cumbersome, however, as the majority of patients with cardiovascular disease or diabetes are already on medication that can potentially affect the measurements.
A study with women only would also be of interest, as gender differences certainly exist. Women were not included in this study because of thermoregulatory changes throughout the menstrual cycle and at menopause, as well as the effects of the menstrual cycle on sleep (Manber and Armitage, 1999; Wright and Badia, 1999) and breathing during sleep (Driver et al., 2005).
More studies looking at well-defined groups of OSA patients stratified based on gender, body mass index and disease status would serve as a step towards personalized medicine for OSA patients (Bell, 2004; Ginsburg et al., 2005) and further understanding of OSA pathophysiology.
This study is the first to show interindividual differences in OSA patients based on sleep-related sweating. OSA patients who experience sleep-related sweating have a lower sleep quality, as measured by proportionally less REM sleep and higher blood pressure than other OSA patients. CPAP treatment causes larger decreases in blood pressure and more increases in REM sleep percentage in these OSA patients than in non-sweaters. Sleep-related sweating can possibly be seen as an added indication for CPAP treatment, as these patients are more likely to have higher blood pressure and less REM sleep than other OSA patients. Finding interindividual differences such as sleep-related sweating in OSA patients and relating the differences to OSA pathophysiology and comorbidities is a step towards much-needed personalized medicine (Bell, 2004; Ginsburg et al., 2005).
We would like to thank our great colleagues at the Sleep Unit of Landspitali University Hospital, especially Magdalena Osk Sigurgunnarsdottir and Atli Josefsson for their assistance throughout the project, the skilful technical assistance of Anna Guðrun Sigurdardottir and Elizabeth Cook, the help of Jennifer Montoya with preparation of figures and we express our gratitude for the grants we received. The project was supported financially by the Icelandic Research Fund (2006–2007), the Icelandic Graduate Research Fund (2006–2007) and the Science Fund of Landspitali University Hospital (2005–2006).