Associations between weight change since 20 years of age ... - Nature

International Journal of Obesity (2009) 33, 1396–1401 & 2009 Macmillan Publishers Limited All rights reserved 0307-0565/09 $32.00 www.nature.com/ijo

ORIGINAL ARTICLE Associations between weight change since 20 years of age and sleep-disordered breathing among male truck drivers R Cui1, T Tanigawa2, H Nakano3, S Sakurai2, K Yamagishi4, T Ohira1 and H Iso1 1 Public Health, Department of Social and Environmental Medicine, Osaka University Graduate School of Medicine, Osaka, Japan; 2Department of Public Health, Doctoral Program in Social Medicine, Graduate School of Medicine, Ehime University, Ehime, Japan; 3Department of Pulmonology, Fukuoka National Hospital, Fukuoka, Japan and 4Department of Public Health Medicine, Graduate School of Comprehensive Human Sciences, Institute of Community Medicine, University of Tsukuba, Ibaraki, Japan

Background: Limited evidence for association of weight gain with sleep-disordered breathing (SDB) has been produced for Asian populations whose body mass index (BMI) levels are lower than in western countries. Objective: The aim of this study was to examine weight change since 20 years of age and risk of SDB among Japanese. Design: Retrospective cohort study. Subjects: This study includes a large sample of 5320 male Japanese truck drivers aged 30–69 years. Measurements: The respiratory disturbance index (RDI) was selected as an indicator of SDB, and it was estimated with a onenight sleep test using an airflow monitor, and the Epworth Sleepiness Scale (ESS) was used to estimate excessive daytime sleepiness. Results: Respiratory disturbance and sleepiness were more prevalent among men with BMI of 25.0–29.9 and X30.0 kg/m2 than among those with BMI of 18.5–24.9; multivariable odds ratios (ORs) were 1.8(1.5–2.0), Po0.001 and 4.4(3.5–5.5), Po0.001 for RDI X10, and 1.2(0.9–1.4), P ¼ 0.18 and 1.5(1.1–2.1), P ¼ 0.02 for ESS X11, respectively. Compared with men showing BMI changes within ±1.0, the respective multivariable ORs for those with BMI changes of 3.0–4.9 and X5.0 were 1.4(1.2–1.6), Po0.001 and 2.4(2.0–2.9), Po0.001 for RDI X10, and 1.2(0.9–1.6), P ¼ 0.22 and 2.0(1.5–2.6), Po0.001 for ESS X11. The corresponding ORs for weight gain of X10.0 kg compared with weight change less than ±5.0 kg were 2.0(1.7–2.4), Po0.001 for RDI X10 and 1.5(1.2–2.0), P ¼ 0.002 for ESS X11. Similar trends were observed for RDI X20. Conclusion: Our results suggest that an increase in BMI of X5 kg/m2 or weight gain of X10 kg is a risk factor for SDB and excessive daytime sleepiness among Japanese truck drivers. International Journal of Obesity (2009) 33, 1396–1401; doi:10.1038/ijo.2009.192; published online 22 September 2009 Keywords: sleep-disordered breathing; weight change; body mass index; epidemiology; Japanese

Introduction Sleep-disordered breathing (SDB) is strongly associated with obesity and overweight,1,2 and with weight gain.3,4 In the USA, more than 80% of the estimated deaths attributable to obesity occurred among individuals with body mass index

Correspondence: Professor T Tanigawa, Department of Public Health, Doctoral Program in Social Medicine, Graduate School of Medicine, Ehime University, Shitsukawa, Toon, Ehime 791-0295, Japan. E-mail: [email protected] Received 31 December 2008; revised 18 August 2009; accepted 23 August 2009; published online 22 September 2009

(BMI) of X30,5 and in 2004 the prevalence of BMI of X30 was 32% among adults aged 20 years or older.6 In Japan, the prevalence of BMI X30 in 2004 was much lower at 3% among adults aged 15 years or older.7 The prevalence of SDB (apnea-hypopnea index (AHI), X15) was 9–14% for western adult men8–10 and 5–10% for their Asian counterparts.11,12 We previously reported that SDB was positively associated with overweight and excessive daytime sleepiness among Japanese adults.13,14 To date, however, no data have been available on the association of weight change with SDB and excessive daytime sleepiness for Asian populations. Our a priori hypothesis was that weight gain as well as overweight would

Weight change and sleep-disordered breathing R Cui et al

1397 be associated with higher prevalence of SDB and excessive daytime sleepiness. In this study, we used the respiratory disturbance index (RDI) as an indicator of SDB and the Epworth Sleepiness Scale (ESS) to assess daytime sleepiness for an investigation of the association of BMI and weight gain with SDB and excessive daytime sleepiness in a large sample of 5320 male Japanese truck drivers.

Methods Subjects From November 2005 to October 2006, 5487 truck drivers aged 30–69 years, all members of the Japanese Trucking Association, participated in this study (participation rate, 97%). We excluded 22 subjects whose RDI data were missing and 145 women because of the small sample size, so that a total of 5320 men were actually enrolled. Physician epidemiologists or trained staff members explained the study protocol to every subject and obtained informed consent. The study was approved by the Medical Ethics Committees of the University of Tsukuba.

Measurements All participants completed self-administered questionnaires about systolic and diastolic blood pressure, current height and weight, weight at the age of 20, ethanol intake per day, number of cigarettes smoked per day, use of antihypertensive medication and the ESS, that is, the sum of the scores from 1 to 3 for eight items (range, 0–24).15 Hypertension was defined as systolic blood pressure X140 mm Hg and/or diastolic blood pressure X90 mm Hg and/or use of antihypertensive medication. BMI was calculated as weight (kg) divided by the square of height in meters (m2), BMI at the age of 20 was estimated by comparing reported weight at the age of 20 with current height on the assumption that height had not changed materially since the age of 20. Persons who smoked one or more cigarettes per day were defined as current smokers. ESS X11 was taken to represent excessive daytime sleepiness as in a previous study.16

Assessment of SDB We asked the participants to use a single-channel airflow monitor (Somnie; NGK Spark Plug Co. Ltd, Nagoya, Japan) with a portable monitor attached below the nose during one night17 of sleep at home and to fill in the self-administered questionnaire on the same day. The Somnie uses a polyvinylidene fluoride film as a thermal sensor to detect airflow, stores the airflow signal as digital data at a sampling frequency of 10 Hz and can record data for 24 h. The airflow sensor is designed to detect both nasal and oral breathing. For the detection of apnea and hypopnea, the Somnie primarily uses the airflow signal obtained by the thermal sensor. The data from the Somnie were analyzed automati-

cally with a computer program (Flow.exe; Institute of Sleep Health Promotion, Tsukuba, Japan), which uses short-time power spectral analysis and yields an index known as the flow-RDI. The strength of this method lies in strong agreement between the automated and manual analyses of airflow signal, which has been confirmed in a previous study.17 The criterion for SDB was defined by the RDI level as 10 and 20 events per hour, because these RDI cutoffs were found to represent AHIs of X15 and X30, respectively, as determined by full polysomnography (PSG); the corresponding sensitivities were 0.91 and 0.89, and specificities 0.82 and 0.96.17 We eventually used a cutoff for RDI of 10 (corresponding to AHI of 15) for the analysis, because the number of RDI X20 (corresponding to AHI of 30) in weight less categories is too small (n ¼ 20).

Statistical analysis Age-adjusted mean values and prevalence of selected cardiovascular risk factors were calculated by using the four BMI categories according to the WHO clinical guidelines (o18.5, 18.5–24.9, 25.0–29.9 and X30.0 kg m2) and subjected to analysis of covariance and the w2-test. We used logistic regression analysis to estimate the independent associations of the four BMI categories, BMI change (p 1.0, within±1.0, 1.0–2.9, 3.0–4.9 and X5.0) and weight change (p 10.0, between 9.9 and 5.0, within±5.0, 5.0–9.9 and X10.0 kg) since the age of 20 with RDI X10 and ESS X11. The various BMI and weight categories were tested for linear trends by using logistic regression models for ordinal variables as references for normal BMI and no BMI or weight changes as well as by using linear regression for continuous variables, which used the median variables of BMI, BMI change and weight change for each of categories. Potential confounding factors for adjustment were age (years), BMI at age 20 (kg m2), weight at age 20 (kg), current regular drinkers (yes), current smokers (yes) and hypertension. SAS version 9.1.3 software (SAS Institute Inc., Cary, NC, USA) was used for all statistical analyses. All probability values for statistical tests were two-tailed, and values of Po0.05 were regarded as statistically significant.

Results Table 1 shows the population characteristics according to the BMI categories. Mean (s.d.) among total subjects were 44.7(9.3) years for age, 24.8(3.6) kg m2 for BMI, 130(14) mm Hg for systolic blood pressure and 69(30) mm Hg for diastolic blood pressure, 9.1(10.5) per hour for RDI and 5.1(3.9) for ESS. For all subjects, the prevalence of current ethanol intake was 54%, current smoking 66%, use of antihypertensive medication 10% and hypertension 30%. Men with higher BMI were younger, more hypertensive, used more medication for hypertension, smoked less, drank less International Journal of Obesity


1398 Table 1

Age-adjusted means (s.e.) and prevalence of risk characteristics for 5320 men according to BMI BMI (kg m2)

Total subjects

Number Age (years) RDI ESS BMI (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Current smokers (%) Current regular drinkers (%) Hypertension (%) Medication use for hypertension (%)

5320 44.7 (9.3) 9.1 (10.5) 5.1 (3.9) 24.8 (3.6) 129.7 (14.0) 69.1 (30.2) 65.7 54.4 30.0 10.4

o18.5

18.5–24.9

25.0–29.9

X30.0

P for trend

106 45.8 (0.9) 5.8 (1.0) 3.8 (0.4) 17.8 (0.18) 124.4 (1.5) 64.0 (3.1) 81.2 50.8 19.3 3.6

2956 44.6 (0.2) 7.2 (0.2) 4.8 (0.1) 22.5 (0.03) 127.1 (0.3) 67.1 (0.6) 67.7 57.8 23.0 6.8

1825 45.3 (0.2) 10.0 (0.2) 5.4 (0.1) 27.0 (0.04) 132.5 (0.3) 71.6 (0.7) 61.2 52.1 37.7 14.4

433 41.8 (0.4) 18.7 (0.5) 5.7 (0.2) 32.7 (0.08) 137.1 (0.7) 74.7 (1.5) 67.5 41.5 48.0 19.5

0.02 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001

Abbreviations: BMI, body mass index; ESS, excessive sleepiness score; RDI, respiratory disturbance index. For ‘Total subjects’ shows the mean and s.d.

Table 2

ORs and 95% CIs of RDI X10 and ESS X11 according to BMI BMI (kg m2) o18.5

18.5–24.9

25.0–29.9

X30.0

106 17.8

2956 22.5

1825 27.0

433 32.7

RDI X10 Number (%) Age-adjusted OR (95% CI) P for differencea Multivariable OR (95% CI) P for differencea

22 (20.8) 0.9 (0.5–1.4) 0.51 0.9 (0.5–1.5) 0.74

665 (22.5) 1.0 F 1.0 F

627 (34.4) 1.8 (1.6–2.0) o0.001 1.8 (1.5–2.0) o0.001

220 (50.8) 4.4 (3.6–5.5) o0.001 4.4 (3.5–5.5) o0.001

ESS X11 Number (%) Age-adjusted OR (95% CI) P for differencea Multivariable OR (95% CI) P for differencea

5 (4.7) 0.6 (0.3–1.6) 0.32 0.7 (0.3–1.8) 0.75

219 (7.4) 1.0 F 1.0 F

159 (8.7) 1.2 (1.0–1.5) 0.08 1.2 (0.9–1.4) 0.18

51 (11.8) 1.6 (1.1–2.2) 0.006 1.5 (1.1–2.1) 0.02

Total number Mean BMI

P for trend

o0.001 o0.001

0.006 0.02

Abbreviations: BMI, body mass index; OR, odds ratio; RDI, respiratory disturbance index. Multivariable adjustment: age (years), current smokers (yes), current regular drinkers (yes) and hypertension (yes). aCompared with BMI of 18.5–24.9 category.

and had higher RDI, ESS and blood pressure levels compared with those with lower BMI. Table 2 shows odds ratios for RDI X10 and ESS X11 by BMI category. The prevalence of RDI X10 and ESS X11 correlated linearly and positively with BMI. Compared to men with BMI of 18.5–24.9, those with BMI of 25.0–29.9 and X30.0 had 1.8- and 4.4-fold higher prevalence of RDI X10, respectively, and those with BMI of X30.0 had 1.5-fold higher prevalence of ESS X11 after adjustment for cardiovascular risk factors. Men with BMI change of X5.0 compared to those with BMI change within±1.0, and men with weight change of X10.0 kg compared to those with weight change within±5.0 kg since the age of 20 showed approximately twice as high a prevalence of RDI X10 and ESS X11 after adjustment for cardiovascular risk factors (Tables 3 and 4). International Journal of Obesity

A 1.5 times higher prevalence of RDI X10 was observed in men with a BMI change of 3.0–4.9. Similar trends were observed for RDI X5 and X20 (not shown in the table).

Discussion In our study of 5320 Japanese truck drivers, we found that men with BMI of 25.0–29.9 and X30.0 had approximately two- and fourfold higher prevalence of SDB (RDI X10), respectively, than those with BMI of 18.5–24.9. Furthermore, men with a BMI change of X5.0 compared with those with a BMI change within±1.0 kg m2, and men with a weight gain of X10.0 kg compared with those with a weight change within±5.0 kg since the age of 20 showed approximately twice as high prevalence of SDB. The risk of SDB also showed


1399 Table 3

ORs and 95% CI of RDI X10 and ESS X11 according to change in BMI BMI change since 20 years of age (kg m2)

P for trend

p1.0

0.9 to 0.9

1.0 to 2.9

3.0 to 4.9

X5.0

344 23.4 26.3

849 21.7 21.5

1605 23.5 21.6

1381 25.4 21.5

1141 28.4 21.4


98 (28.5) 1.2 (0.9–1.6) 0.21 0.9 (0.7–1.2) 0.60

185 (21.8) 1.0 F 1.0 F

363 (22.6) 0.9 (0.8–1.1) 0.53 1.0 (0.8–1.2) 0.74

407 (29.5) 1.4 (1.2–1.6) o0.001 1.4 (1.2–1.6) o0.001

481 (42.2) 2.4 (2.0–2.8) o0.001 2.4 (2.0–2.9) o0.001


15 (4.4) 0.7 (0.4–1.2) 0.17 0.7 (0.4–1.4) 0.32

60 (7.1) 1.0 F 1.0 F

107 (6.7) 1.1 (0.8–1.5) 0.69 1.1 (0.8–1.5) 0.68

109 (7.9) 1.2 (0.9–1.6) 0.13 1.2 (0.9–1.6) 0.22

143 (12.5) 2.1 (1.6–2.7) o0.001 2.0 (1.5–2.6) o0.001

Total number Mean BMI Mean BMI at age 20 years

o0.001 o0.001

o0.001 o0.001

Abbreviations: BMI, body mass index; OR, odds ratio; RDI, respiratory disturbance index. Multivariable adjustment variables are shown in Table 2 and further for BMI (kg m2) at 20 years of age. aCompared with BMI change within±1.0 category.

Table 4

ORs and 95% CI of RDI X10 and ESS X11 according to weight change Weight change since 20 years of age (kg)

P for trend

p10.0

9.9 to 5.0

4.9 to 4.9

5.0 to 9.9

X10.0

111 25.4 30.2

130 22.8 25.1

1346 22.0 21.6

1413 24.0 21.6

2320 26.9 21.4


31 (27.9) 1.4 (0.9–2.2) 0.13 0.7 (0.4–1.1) 0.15

38 (29.2) 1.3 (0.9–2.0) 0.19 1.1 (0.7–1.7) 0.66

295 (21.9) 1.0 F 1.0 F

341 (24.1) 1.2 (1.0–1.4) 0.10 1.1 (0.9–1.4) 0.16

829 (35.7) 2.0 (1.7–2.3) o0.001 2.0 (1.7–2.4) o0.001


3 (2.7) 0.4 (0.1–1.2) 0.11 0.4 (0.1–1.4) 0.15

4 (3.6) 0.5 (0.2–1.3) 0.14 0.5 (0.2–1.4) 0.19

90 (6.7) 1.0 F 1.0 F

101 (7.1) 1.1 (0.8–1.4) 0.66 1.0 (0.8–1.4) 0.79

236 (10.2) 1.6 (1.2–2.1) o0.001 1.5 (1.2–2.0) 0.002

Total number Mean BMI Mean BMI at age 20 years

o0.001 o0.001

o0.001 0.002

Abbreviations: BMI, body mass index; OR, odds ratio; RDI, respiratory disturbance index. Multivariable adjustment variables are shown in Table 2 and further for weight at 20 years of age. aCompared with weight change within±5.0 category.

a 1.5-fold increase even for smaller weight gains such as a BMI increase of 3.0–4.9. These strongly positive associations of BMI and BMI changes since the age of 20 with SDB were consistent with the results of previous studies, but this study provided further evidence that long-term changes in BMI are associated with SDB.10,12–14 In our study, 55% of the men with RDI X10 had BMI X25, which is similar to findings for the USA, where 58% of Americans with AHI X15 had BMI X25.1 A 10% increase in weight over 4 years was associated with a six times higher risk of AHI X153 and a weight gain of X5 kg

over 5 years with a three times higher risk of AHI X15 18 in follow-up studies of Americans. We found that prevalence of SDB was elevated for BMI X25.0 and for BMI gain of X3 and weight gain of X10 kg since the age of 20. A recent 10-year follow-up study of Japanese subjects reported that weight gain of X10 kg among men with BMI o21.7 at age 20 was associated with a doubling of the risk of coronary heart disease but that no excessive risk was established for men with BMI X21.7.19 We also found that the prevalence of excessive daytime sleepiness was elevated for BMI X30 and for BMI gain of X3 International Journal of Obesity


1400 and weight gain of X10 kg since the age of 20. Previous epidemiological studies showed that excessive daytime sleepiness was positively associated with SDB 8,9,13 and with risk of traffic accidents,20 and that the prevalence of SDB and sleepiness tended to be lower at weight loss of X10 kg. Putting these findings together suggests that weight control for professional drivers may be beneficial for the prevention of SDB. The strength of this study is that it is the first to examine SDB status during sleep at home for a large sample of professional drivers, which has the advantage of providing a more realistic estimation of severity of SDB compared with hospital/laboratory studies because the subjects can maintain regular daily habits of sleep, physical activity, diet and ethanol intake. Our study has several limitations, however. First, we used the RDI but not 3% oxygen desaturation index (ODI) as an indicator of SDB, because the sensitivity of 3% ODI for the diagnosis of SDB was not high for nonobese persons. The sensitivity of 3% ODI X5 for screening for AHI X5 by means of full PSG was 68% for subjects with BMI p27.0.21 However, the sensitivity of RDI X5 for the same screening was 97% for subjects with BMI o25.0.17 Second, we conducted airflow monitoring for all participants only once, which is likely to have resulted in underestimation of associations between weight change and SDB due to measurement errors. Because the misclassification of airflow findings is nondifferential among weight categories, a real association world is stronger. Third, we used self-reported weight and height to calculate BMI, and these measurements may contain errors in the form of underestimated weight and overestimated height.22 However, the validity for BMI estimated from self-reported height and weight is reportedly very good for Japanese (r ¼ 0.94).23 Fourth, we used recall data for weight at the age of 20, which may have been inaccurate. In Japan, however, annual medical checkups have been conducted since 1958 for students and employees at around 20 years of age under the School Health Law and since 1972 under the Occupational Safety and Health Act, which may help subjects to recall their weight more accurately.24 Another study reported that recalled weight at the age of 25 strongly correlated with actually measured weight (r ¼ 0.85).25 Finally, because our study population was limited to male truck drivers, it is not clear whether our results can be extrapolated to female truck drivers or the general population. In conclusion, significant associations of weight gain with SDB and excessive daytime sleepiness suggest the need for avoidance of weight again among professional drivers to prevent SDB.

Acknowledgements This study was supported in part by Grants-in-Aid for Scientific Research B (No. 14370132 in 2002–2005) from International Journal of Obesity

the Japan Society for the Promotion of Science, and by the Sprout Research (No. 17659184 in 2005–2006) from the Ministry of Education, Culture, Sports, Science and Technology, Japan. We are grateful to Mr Akimoto Yutaka of the Japanese Trucking Association for his helpful assistance.

References 1 Young T, Peppard PE, Taheri S. Excess weight and sleep-disordered breathing. J Appl Physiol 2005; 99: 1592–1599. 2 Young T, Shahar E, Nieto FJ, Redline S, Newman AB, Gottlieb DJ et al. Predictors of sleep-disordered breathing in communitydwelling adults: the Sleep Heart Health Study. Arch Intern Med 2002; 162: 893–900. 3 Peppard PE, Young T, Palta M, Dempsey J, Skatrud J. Longitudinal study of moderate weight change and sleep-disordered breathing. JAMA 2000; 284: 3015–3021. 4 Redline S, Schluchter MD, Larkin EK, Tishler PV. Predictors of longitudinal change in sleep-disordered breathing in a nonclinic population. Sleep 2003; 26: 703–709. 5 Allison DB, Fontaine KR, Manson JE, Stevens J, VanItallie TB. Annual deaths attributable to obesity in the United States. JAMA 1999; 282: 1530–1538. 6 Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM. Prevalence of overweight and obesity in the United States, 1999–2004. JAMA 2006; 295: 1549–1555. 7 Ministry of Health Labour and Welfare Japan. The National Health and Nutrition Survey in Japan, 2004. Daiichi Shuppan: Tokyo, 2006. (in Japanese). 8 Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993; 328: 1230–1235. 9 Gottlieb DJ, Whitney CW, Bonekat WH, Iber C, James GD, Lebowitz M et al. Relation of sleepiness to respiratory disturbance index: the Sleep Heart Health Study. Am J Respir Crit Care Med 1999; 159: 502–507. 10 Dura´n J, Esnaola S, Rubio R, Iztueta A. Obstructive sleep apneahypopnea and related clinical features in a population-based sample of subjects aged 30–70 yr. Am J Respir Crit Care Med 2001; 163: 685–689. 11 Ip MS, Lam B, Lauder IJ, Tsang KW, Chung KF, Mok YW et al. A community study of sleep-disordered breathing in middle-aged Chinese men in Hong Kong. Chest 2001; 119: 62–69. 12 Kim J, In K, Kim J, You S, Kang K, Shim J et al. Prevalence of sleepdisordered breathing in middle-aged Korean men and women. Am J Respir Crit Care Med 2004; 170: 1108–1113. 13 Tanigawa T, Tachibana N, Yamagishi K, Muraki I, Kudo M, Ohira T et al. Relationship between sleep-disordered breathing and blood pressure levels in community-based samples of Japanese men. Hypertens Res 2004; 27: 479–484. 14 Cui R, Tanigawa T, Sakurai S, Yamagishi K, Iso H. Relationships between sleep-disordered breathing and blood pressure and excessive daytime sleepiness among truck drivers. Hypertens Res 2006; 29: 605–610. 15 Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 1991; 14: 540–545. 16 Johns M, Hocking B. Daytime sleepiness and sleep habits of Australian workers. Sleep 1997; 20: 844–849. 17 Nakano H, Tanigawa T, Ohnishi Y, Uemori H, Senzaki K, Furukawa T et al. Validation of a single-channel airflow monitor for screening of sleep-disordered breathing. Eur Respir J 2008; 32: 1060–1067. 18 Newman AB, Foster G, Givelber R, Nieto FJ, Redline S, Young T. Progression and regression of sleep-disordered breathing with changes in weight: the Sleep Heart Health Study. Arch Intern Med 2005; 165: 2408–2413.


1401 19 Chei CL, Iso H, Yamagishi K, Inoue M, Tsugane S. Body mass index and weight change since 20 years of age and risk of coronary heart disease among Japanese: the Japan Public Health Center-Based Study. Int J Obes 2008; 32: 144–151. 20 Howard ME, Desai AV, Grunstein RR, Hukins C, Armstrong JG, Joffe D et al. Sleepiness, sleep-disordered breathing, and accident risk factors in commercial vehicle drivers. Am J Respir Crit Care Med 2004; 170: 1014–1021. 21 Nakamata M, Kubota Y, Sakai K, Kinebuchi S, Nakayama H, Oodaira T, Sato M, Shinoda H, Kouno M. The limitation of screening test for patients with sleep apnea syndrome using pulse oximetry. J Jap Soc Respir Care 2003; 12: 401–406 (in Japanese).

22 Stevens J, Keil JE, Waid LR, Gazes PC. Accuracy of current, 4-year, and 28-year self-reported body weight in an elderly population. Am J Epidemiol 1990; 132: 1156–1163. 23 Kawada T, Takeuchi K, Suzuki S, Aoki S. Association of reported and recorded height, weight and blood pressure. Nipponn Koshu Eisei Zasshi 1994; 41: 1099–1103 (in Japanese). 24 Health Welfare Statistics Association. Kokumin eisei no doko. J Health Welf Stat 2008; 55: 303–374 (in Japanese). 25 Tamakoshi K, Yatsuya H, Kondo T, Hirano T, Hori Y, Yoshida T et al. The accuracy of long-term recall of past body weight in Japanese adult men. Int J Obes Relat Metab Disord 2003; 27: 247–252.

International Journal of Obesity