Combined effects of smoking and peripheral arterial disease on all-cause and cardiovascular disease mortality in a Chinese male cohort

Transcription

1 Combined effects of smoking and peripheral arterial disease on all-cause and cardiovascular disease mortality in a Chinese male cohort Yingyi Luo, MD, a,b Xiankai Li, MD, b Jue Li, MD, PhD, b Xiaoming Wang, MD, a Yongxia Qiao, MD, b Dayi Hu, MD, b Philip A. Merriam, MSPH, c and Yunsheng Ma, MD, PhD, c Shanghai, China; and Worcester, Mass Objective: Smoking is a major risk factor for peripheral arterial disease (PAD), and PAD is associated with all-cause and cardiovascular disease (CVD) mortality. The objective of this study was to determine the combined effects of smoking and PAD on all-cause and CVD mortality. Methods: A total of 1979 males 35 years of age or older were enrolled from eight university-affiliated hospitals in Beijing and Shanghai in 2004, with both smoking status and PAD diagnosis obtained, 1712 of them had complete follow-up data. Mortality data were obtained from all participants between December 2007 and February Cox proportional hazards models were used to evaluate relative risks (RRs) of all-cause mortality and CVD mortality among different groups. Results: At baseline, the average age of participants was years-old (SD 11.57), prevalence of PAD was 24.0% and 65.4% smoked cigarettes. During the 3-year follow-up, all-cause cumulative mortality rates were 27.9% (PAD/smoker), 26.3% (PAD/nonsmoker), 14.1% (no PAD/smoker), and 14.4% (no PAD/nonsmoker) (P <.001), and CVD cumulative mortality rates were 17.8%, 14.9%, 8.1%, and 7.3%, respectively (P <.001). Compared with the no PAD/nonsmoker subjects, adjusted RR from all-cause mortality in the groups of both PAD/smoker, PAD/nonsmoker, and no PAD/ smoker were 1.88 (95% confidence interval [CI], ), 1.37 (95% CI, ), and 1.08 (95% CI, ), respectively. The adjusted RR from CVD mortality was 2.12 (95% CI, ), 1.55 (95% CI, ), and 1.13 (95% CI, ), respectively. Conclusion: PAD is a major determinant of mortality. Smoking did not contribute to mortality in this study. Further research is needed. (J Vasc Surg 2010;51:673-8.) From the Shanghai Institute of Health Sciences, a Heart, Lung, and Blood Vessel Center, Tongji University School of Medicine, b and the Division of Preventive and Behavioral Medicine, Department of Medicine, University of Massachusetts Medical School. c Competition of interest: none. Reprint requests: Jue Li, MD, PhD, Tongji University School of Medicine, Heart, Lung, and Blood Vessel Research Center, 1239 Siping Road, Shanghai , China ( jueli59jp@yahoo.co.jp). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a competition of interest /$36.00 Copyright 2010 Published by Elsevier Inc. on behalf of the Society for Vascular Surgery. doi: /j.jvs Peripheral arterial disease (PAD) is a clinical manifestation of the atherosclerotic process, which is associated with the increased prevalence of cardiovascular disease (CVD) and elevated CVD mortality. 1,2 Many studies have shown that PAD is a strong prospective predictor of CVD mortality and all-cause mortality, independent of other adjusted risk factors. 3-6 Up to 20% of elderly individuals have PAD upon noninvasive testing, 7-10 and is common in elderly Chinese people. 11 Because only a small percentage of these individuals are symptomatic, PAD is difficult to be recognized in primary care practice. 12,13 In China, He et al 11 showed that 43% of the PAD patients were asymptomatic and unaware of their condition. As a consequence, numerous studies have attempted to assist physicians in identifying individuals with the greatest risk of developing PAD. Such studies consistently identify smoking as a key risk factor. 14,15 Because smoking is also an important predictor of mortality, 16,17 it was hypothesized that the risk of death is higher in smoking patients with PAD than in nonsmoking patients with PAD. Except for a couple of 1-year studies that reported the relationship between PAD and mortality in Chinese patients, 18,19 this issue has not been researched extensively in China. The prevalence of smoking in China is as high as 59.3% in males whereas few females smoke. 20 Estimates from 1988 showed that China consumes 30% of all cigarettes worldwide (ie, 1.5 trillion cigarettes). 21 It has been predicted that in China alone, 50 million individuals currently under 20 years of age will ultimately die of tobaccorelated diseases. 22 The combined effect of smoking and PAD to all-cause and CVD mortality in Chinese male adults is not clear. Thus, the purposes of this study were to investigate the association between smoking and PAD, and the risk of all-cause and CVD mortality for smoking and nonsmoking Chinese males with and without PAD. MATERIALS AND METHODS Study subjects. The ankle brachial index (ABI) Cohort Study was designed in 2004 to investigate the risk factors of PAD and the relationship between ABI and 673

2 674 Luo et al JOURNAL OF VASCULAR SURGERY March 2010 mortality from all-cause and CVD. Subjects were recruited from the department of endocrinology or a cardiology in-patient clinic located in eight university-attached hospitals located in Beijing and Shanghai between July and November A total of 1979 male participants, 35 years or older who had complete baseline data, were entered in this study cohort. Follow-up visits were held between December 2007 and February Of the 1979 male participants, 1712 of them had complete follow-up data. This study was approved by the ethics committee of Tongji University and informed consent was obtained from each participant. Measurement of ankle and arm blood pressures. Doppler ultrasound scan (Nicolet Vascular, Elite 100R, Madison, Wisc) was used to measure systolic pressure on bilateral brachial, posterior tibial, and dorsal pedal arteries in the supine position after a 5-minute rest. The occluding cuffs ( cm) were applied just above the malleoli to measure ankle pressure. The Doppler scan probe was used at a frequency of 5 MHz. Right and left ABIs were calculated by the highest pressure on the dorsal or posterior tibial arteries on the right and left sides, respectively, and by the highest brachial pressure on either side. Diagnosis of peripheral arterial disease. The American College of Cardiology and the American Heart Association (ACC/AHA) guidelines have recommended ABI 0.90 as the criterion for the diagnosis of PAD. 23 Our study adopted this recommendation for participants with an ABI 0.90 in either leg as PAD, whereas participants with an ABI from 0.91 to 1.40 were considered not to have PAD. Classification of smoking behavior. Participants were asked to describe their smoking habits in the 2004 baseline questionnaire, including the number of cigarettes smoked per day and the duration of cigarette smoking. Smoker refers to the participants who reported smoking cigarettes or had smoked other tobacco products at least once a day for 1 year or more. 24 The dose of smoking was further classified by the average number of cigarettes smoked per day (0, 1-15, and 16 cigarettes per day) and by the cumulative number of cigarettes ever smoked (0, 365, 100,000, 200,000, 300,000, 400,000 cigarettes). Diagnosis of diabetes, hypertension, chronic kidney disease, and coronary heart disease. Diabetes was defined as having a fasting glucose 7 mmol/l, reported use of hypoglycemic medication, or reported diabetes at the baseline examination; hypertension: a systolic blood pressure (SBP) 140 mm Hg and/or diastolic blood pressure (DBP) 90 mm Hg, reported use of hypertensive medications, or reported hypertension history at the baseline examination; chronic kidney disease (CKD): an estimated glomerular filtration rate (egfr) 60 ml/minute per 1.73 m 2 or reported CKD history at the baseline examination; and coronary heart disease (CHD): confirmed by coronariography, a previous myocardial infarction, or surgery or coronary revascularization. All these diagnoses for participants were reviewed by study physicians. Table I. The proportion of cause of death for male participants of ABI cohort study in China, 3-years followup (n 1712) Cause of death n Proportion (%) Coronary heart disease Ischemic stroke Hemorrhagic stroke Other cardiovascular disease Others Total ABI, Ankle brachial index. Identification of all-cause and cardiovascular disease mortality. Death was confirmed from hospital records or by contact with participants and their families. Cause of death was investigated using medical records and key informant interviews. The CVD mortality was estimated according to the International Classification of Diseases 10 (ICD-10). All materials were reviewed independently and blindly by a physician to confirm the cause of death. Statistical analysis. Data were analyzed using the software program SPSS 13.0 (SPSS Inc, Chicago, Ill). Continuous variables were expressed as the mean SD, and categoric variables as a percentage. Independent samples t test and the 2 test were used to compare continuous and categoric variables, respectively. The cumulative mortality rates from all-cause and CVD were calculated by different groups. Survival analysis based on Kaplan-Meier curves and log-rank tests were used to compare survival rates in different smoker/pad groups. The relative risks of death from all-cause and CVD were adjusted for potential confounders including age, systolic blood pressure (SBP), body mass index (BMI), diabetes, hypertension, stroke, coronary heart disease, and chronic kidney disease using Cox proportional regression models. RESULTS There were 1712 male participants whose baseline data and follow-up data was available. Among them, 411 participants (24.0%) were in the low-abi group and diagnosed with PAD, with the remainder in the normal-abi group. According to our criterion described above, the prevalence of PAD was 24.0% in this cohort. The prevalence of smoking was 65.4%. During a mean follow-up time of months, 267 (13.5%) of them were lost to follow-up because of changing telephone numbers and family addresses. There is no significant difference in the smoking rates and PAD prevalence between subjects with and without follow-up data. Thus, among the 1712 participants who have complete follow-up data, 298 of them died, with 172 attributed to CVD. The 3-year cumulative mortality rate was 17.4%. The detailed cause of death proportions are shown in Table I.

3 JOURNAL OF VASCULAR SURGERY Volume 51, Number 3 Luo et al 675 Table II. Baseline characteristics for male participants with and without PAD (n 1712) Groups Baseline Characteristics With PAD n 411 Without PAD n 1301 P values Age BMI (kg/m 2 ) SBP (mm Hg) DBP (mm Hg) TC (mmol/l) TG (mmol/l) LDL-C (mmol/l) HDL-C (mmol/l) FPG (mmol/l) Diabetes (%) Dislipidemia (%) Stroke (%) Hypertension (%) Coronary heart disease (%) Chronic kidney disease (%) Smoking history (%) PAD, Peripheral arterial disease; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; FPG, fasting plasma glucose; TC, total cholesterol; TG, triglycerides; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol. Data are means for continuous variables and percents for categorical variables. Comparison of baseline characteristics between groups with and without PAD. As shown in Table II, participants with PAD were older and had a lower BMI, higher SBP, and had a higher rate of diabetes, hypertension, stroke, coronary heart disease, and a history of smoking as compared with those without PAD at baseline. Relationship of smoking and peripheral arterial disease. The prevalence of PAD in smokers and nonsmokers was 26.5% and 19.3%, respectively (P.001). When the dose of smoking was classified into three groups by the number of cigarettes smoked per day (0, 1-15, and 16 cigarettes per day), the prevalence of PAD is 19.3%, 27.5%, and 25.8%, respectively, in each group (P.003). Participants who smoked 1-15 cigarettes per day had the highest PAD prevalence. When the dose of smoking was classified into six groups by cumulative number of cigarettes ever smoked (0, 365, 100,000, 200,000, 300,000, 400,000 cigarettes), the prevalence of PAD is 19.3%, 24.0%, 23.1%, 24.7%, 39.2%, and 31.1%, respectively (P.001). The prevalence of PAD increased sharply when the cumulative number of cigarettes was 300,000. All-cause and cardiovascular disease mortality in different smoking/peripheral arterial disease groups. Fig 1 shows the all-cause and CVD mortality in the four different smoking/nonsmoking and PAD/no PAD groups. The 1712 male participants who had complete follow-up data were classified into four groups. A total of 297 of them were both smokers and had PAD, 114 had PAD only, 823 were smokers only, and 478 neither smoked nor had PAD. Fig 1. All-cause and cardiovascular disease (CVD) mortality according to peripheral arterial disease (PAD)/smoker group, the ankle-brachial index (ABI) Cohort Study, 3-years follow-up, (n 1712). The cumulative all-cause mortality was 27.9%, 26.3%, 14.1%, and 14.4%, respectively (P.001), and CVD mortality was 17.8%, 14.9%, 8.1%, and 7.3%, respectively (P.001). Survival analysis for different smoking/peripheral arterial disease groups. Kaplan-Meier curves in Fig 2, a and b show the survival distribution of the four groups for all-cause and CVD mortality, respectively. The survival rate for both PAD and smoker group was significantly lower than that for the no PAD/nonsmoker group for both all-cause and CVD mortality (log-rank: P.001). We compared the survival rates in smokers vs nonsmokers within the PAD cohort. There is no significant difference between the two groups (P.41). Mortality risks according to different smoking/ peripheral arterial disease groups. The four groups were examined in relation to mortality in Cox regression models that adjusted for age, BMI, SBP, history of diabetes, hypertension, stroke, CHD, and CKD. Compared with the group of neither PAD/smoker, the adjusted relative risk (RR) from all-cause mortality in the PAD/smoker group, PAD/nonsmoker group, and no PAD/smoker were 1.88 (95% CI, ), 1.37 (95% CI, ), and 1.08 (95% CI, ), respectively. The adjusted RR from CVD mortality was 2.12 (95% CI, ), 1.55 (95% CI, ), and 1.13 (95% CI, ), respectively. DISCUSSION In this study, PAD prevalence was 24.0% and 65.4% of participants who smoked cigarettes in a Chinese male population. Of all the risk factors of PAD, smoking is one of the most important. Two studies have shown that cigarette smoking can result in a sevenfold increase in the risk of PAD. 25,26 In this study, the baseline characteristics are compared in Table II. The smoking rate in participants with PAD was significantly higher than that in participants without PAD. The prevalence of PAD in smokers was higher than that of nonsmokers (26.5% vs 19.3%; P.001). To

4 676 Luo et al JOURNAL OF VASCULAR SURGERY March 2010 a b Fig 2. a, Kaplan-Meier curves for subjects of peripheral arterial disease (PAD)/smoker groups from all-cause mortality, the ankle-brachial index (ABI) Cohort Study, 3-years follow-up. b, Kaplan-Meier curves for subjects of PAD/smoker groups from CVD mortality, the ABI Cohort Study, 3-years follow-up. assess the risk of PAD according to the dose of smoking, subsequent analysis was performed on subjects grouped according to their cigarettes smoked per day, and lifetime cumulative values. The prevalence of PAD increased with the cumulative number of cigarettes ever smoked, especially when the cumulative number is 300,000 cigarettes. Therefore, to stop smoking is an important way to decrease the PAD prevalence in China s high-risk male population. Smoking is related to mortality in many cohort studies. 16,27,28 A prospective male cohort study has shown that cigarette smoking was an important predictor of risk of mortality in men from Shanghai, and 21% of all-cause mortality could be attributed to cigarette smoking. 16 PAD is also related to mortality. 18 A cardiovascular medicine review about the relationship between PAD and mortality by Golomb et al 3 has shown that most studies demon-

5 JOURNAL OF VASCULAR SURGERY Volume 51, Number 3 Luo et al 677 strated PAD can powerfully predict overall mortality in men and women. This has also been shown in community cohorts, 10,29 medical-based cohorts, 30,31 and in a population at heightened CVD risk. 32 The RR ranged from 1.2 to While this issue is seldom discussed in Asian populations, the combined effect of smoking and PAD to all-cause and CVD mortality in individuals is not clear. Participants were further classified into four groups in this 3-year cohort study; the all-cause mortality in both PAD/smoker was twice that of the no PAD/nonsmoker group, and also higher than in the no PAD/smoker and the only PAD/ nonsmoker group. The CVD mortality in PAD/smoker was nearly three times that of the no PAD/nonsmoker group, and also higher than the no PAD/smoker and PAD/nonsmoker groups. It is possible that smoking will result in PAD, and then increase mortality, especially the mortality attributed to CVD. It is also possible that there is a combined effect of smoking and PAD to all-cause and CVD mortality. In this study, the participants in the PAD/smoker group had the lowest survival rate from survival analysis based on Kaplan-Meier curves. However, the survival rate for nonsmokers is not statistically significantly higher than smokers within the PAD cohort (log-rank, P.41). Although this is counter-intuitive and probably a type 2 error, further investigation is needed. The adjusted RRs of the PAD/smoker group from all-cause and CVD mortality were about twice that of no PAD/smoker or PAD/nonsmoker group, and also higher in the no PAD/smoker and PAD/nonsmoker groups. Thus if a male has a smoking habit and at the same time has PAD, he will be at the highest fatality risk, especially fatality due to CVD. It is hypothesized that atherosclerosis is aggravated because of smoking, which makes mortality, especially CVD mortality increase. Therefore, our suggestion is that male PAD patients should avoid smoking and those who smoke should stop smoking. We suggest that the Chinese government should spend more resources to deal with smoking-related diseases. This is the first study of a large cohort between PAD and mortality in China. Although the subjects are high-risk male participants, it really indicated a combined effect of smoking and PAD to all-cause and CVD mortality. Because the patients came from not only urban but also rural areas, the result will hold for all of the country including rural areas. The relationship between PAD and mortality in the general Chinese population should be considered in the future. A potential limitation of our study is the relatively brief duration of follow-up. Further studies should investigate the pathophysiologic mechanisms in which PAD, combined with a smoking habit, may result in all-cause and CVD mortality. In conclusion, our study analyzed the relationship between smoking and PAD and the risk of all-cause and CVD mortality for males with and without PAD who did and did not smoke, results suggest that PAD is a major determinant of mortality, whereas smoking did not contribute to mortality in this study, further research is needed. This study was financially supported by Shanghai Educational Development Foundation (China). AUTHOR CONTRIBUTIONS Conception and design: YL, XL, JL Analysis and interpretation: YL, XL, JL, XW, YM Data collection: YL, XW, YQ Writing the article: YL, YQ Critical revision of the article: XL, JL, XW, YQ, DH, PM, YM Final approval of the article: YL, XL, JL, XW, YQ, DH, PM, YM Statistical analysis: XL, JL, YM Obtained funding: JL, DH Overall responsibility: JL REFERENCES 1. Hiatt WR. Medical treatment of peripheral arterial disease and claudication. N Engl J Med 2001;344: Murabito JM, Evans JC, Nieto K, Larson MG, Levy D, Wilson PW. Prevalence and clinical correlates of peripheral arterial disease in the Framingham Offspring Study. Am Heart J 2002;143: Golomb BA, Dang TT, Criqui MH. Peripheral arterial disease: morbidity and mortality implications. Circulation 2006;114: Resnick HE, Lindsay RS, McDermott MM, Devereux RB, Jones KL, Fabsitz RR, Howard BV. Relationship of high and low ankle brachial index to all-cause and cardiovascular disease mortality: the Strong Heart Study. Circulation 2004;109: Leibson CL, Ransom JE, Olson W, Zimmerman BR, O Fallon WM, Palumbo PJ. Peripheral arterial disease, diabetes, and mortality. Diabetes Care 2004;27: Sutton-Tyrrell K, Venkitachalam L, Kanaya AM, Boudreau R, Harris T, Thompson T, et al. Relationship of ankle blood pressures to cardiovascular events in older adults. Stroke 2008;39: Novo S. Classification, epidemiology, risk factors, and natural history of peripheral arterial disease. Diabetes Obes Metab 2002;4 Suppl 2:S Meijer WT, Hoes AW, Rutgers D, Bots ML, Hofman A, Grobbee DE. Peripheral arterial disease in the elderly: the Rotterdam Study. Arterioscler Thromb Vasc Biol 1998;18: Stoffers HE, Rinkens PE, Kester AD, Kaiser V, Knottnerus JA. The prevalence of asymptomatic and unrecognized peripheral arterial occlusive disease. Int J Epidemiol 1996;25: Newman AB, Siscovick DS, Manolio TA, Polak J, Fried LP, Borhani NO, Wolfson SK. Ankle-arm index as a marker of atherosclerosis in the Cardiovascular Health Study. Cardiovascular Heart Study (CHS) Collaborative Research Group. Circulation 1993;88: He Y, Jiang Y, Wang J, Fan L, Li X, Hu FB. Prevalence of peripheral arterial disease and its association with smoking in a population-based study in Beijing, China. J Vasc Surg 2006;44: Hirsch AT, Criqui MH, Treat-Jacobson D, Regensteiner JG, Creager MA, Olin JW, et al. Peripheral arterial disease detection, awareness, and treatment in primary care. JAMA 2001;286: McDermott MM, Kerwin DR, Liu K, Martin GJ, O Brien E, Kaplan H, Greenland P. Prevalence and significance of unrecognized lower extremity peripheral arterial disease in general medicine practice. J Gen Intern Med 2001;16: Aronow WS. Peripheral arterial disease in the elderly. Clin Interv Aging 2007;2: Sritara P, Sritara C, Woodward M, Wangsuphachart S, Barzi F, Hengprasith B, Yipintsoi T. Prevalence and risk factors of peripheral arterial disease in a selected Thai population. Angiology 2007;58: Yuan JM, Ross RK, Wang XL, Gao YT, Henderson BE, Yu MC. Morbidity and mortality in relation to cigarette smoking in Shanghai, China. A prospective male cohort study. JAMA 1996;275:

6 678 Luo et al JOURNAL OF VASCULAR SURGERY March Kelly TN, Gu D, Chen J, Huang JF, Chen JC, Duan X, et al. Cigarette smoking and risk of stroke in the chinese adult population. Stroke 2008;39: Luo YY, Li J, Xin Y, Zheng LQ, Yu JM, Hu DY. Risk factors of peripheral arterial disease and relationship between low ankle brachial index and mortality from all-cause and cardiovascular disease in Chinese patients with hypertension. J Hum Hypertens 2007;21: Xu Y, Li J, Luo Y, Wu Y, Zheng L, Yu J, et al. The association between ankle-brachial index and cardiovascular or all-cause mortality in metabolic syndrome of elderly Chinese. Hypertens Res 2007;30: Ministry of Public Health. Smoking prevalence rate in China. P.R. China Tobacco. Cigarette sales are on the up. Tobacco Sept, Peto R. Future mortality from tobacco in China. Paper presented at: Shangai Symposium on Smoking and Health; November, 1987; Shangai, China. 23. Hirsch AT, Haskal ZJ, Hertzer NR, Bakal CW, Creager MA, Halperin JL, et al. ACC/AHA 2005 Practice Guidelines for the management of patients with peripheral arterial disease (lower extremity, renal, mesenteric, and abdominal aortic): a collaborative report from the American Association for Vascular Surgery/Society for Vascular Surgery, Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, Society of Interventional Radiology, and the ACC/AHA Task Force on Practice Guidelines (Writing Committee to Develop Guidelines for the Management of Patients With Peripheral Arterial Disease): endorsed by the American Association of Cardiovascular and Pulmonary Rehabilitation; National Heart, Lung, and Blood Institute; Society for Vascular Nursing; TransAtlantic Inter-Society Consensus; and Vascular Disease Foundation. Circulation 2006;113: e WHO. Arterial hypertension report of a WHO expert committee Geneva, 1984, WHO. 25. Heliövaara M, Karvonen MJ, Vilhunen R, Punsar S. Smoking, carbon monoxide, and atherosclerotic diseases. Br Med J 1978;1: Hughson WG, Mann JI, Tibbs DJ, Woods HF, Walton I. Intermittent claudication: factors determining outcome. Br Med J 1978;1: Jacobs DR Jr, Adachi H, Mulder I, Kromhout D, Menotti A, Nissinen A, Blackburn H. Cigarette smoking and mortality risk: twenty-five-year follow-up of the Seven Countries Study. Arch Intern Med 1999;159: Pham TM, Fujino Y, Ide R, Shirane K, Tokui N, Kubo T, et al. Mortality attributable to cigarette smoking in a cohort study in Japan. Eur J Epidemiol 2007;22: Criqui MH, Langer RD, Fronek A, Feigelson HS, Klauber MR, McCann TJ, Browner D. Mortality over a period of 10 years in patients with peripheral arterial disease. N Engl J Med 1992;326: Leng GC, Fowkes FG, Lee AJ, Dunbar J, Housley E, Ruckley CV. Use of ankle brachial pressure index to predict cardiovascular events and death: a cohort study. BMJ 1996;313: Behar S, Zion M, Reicher-Reiss H, Kaplinsky E, Goldbourt U. Shortand long-term prognosis of patients with a first acute myocardial infarction with concomitant peripheral vascular disease. SPRINT Study Group. Am J Med 1994;96: Ogren M, Hedblad B, Engström G, Janzon L. Prevalence and prognostic significance of asymptomatic peripheral arterial disease in 68- year-old men with diabetes. Results from the population study Men born in 1914 from Malmö, Sweden. Eur J Vasc Endovasc Surg 2005;29: Submitted Jul 16, 2009; accepted Sep 12, 2009.