Kidney International, Vol. 62 (2002), pp. 956–962

Influence of smoking and obesity on the developmentof proteinuria

MASAHIKO TOZAWA, KUNITOSHI ISEKI, CHIHO ISEKI, SAORI OSHIRO, YOSHIHARU IKEMIYA, andSHUICHI TAKISHITA

Third Department of Internal Medicine and Dialysis Unit, University of The Ryukyus, Okinawa General Health MaintenanceAssociation, Okinawa, Japan

teinuria independently of both hypertension and diabetes mel-Influence of smoking and obesity on the development of pro-litus in men. The risk of developing proteinuria also tendedteinuria.to be increased with cigarette smoking in men. Smoking andBackground. Proteinuria is a significant risk factor for end-obesity in women were not significant in this data set.stage renal disease. Previous evidence suggested that smoking

and obesity increase the risk of proteinuria. However, it isunclear whether these risk factors predict the development ofproteinuria independently of hypertension and diabetes melli- Proteinuria was identified as a significant predictor oftus. The aim of this study was to analyze the effects of obesity

end-stage renal disease (ESRD) in a mass screening ofand smoking on the development of proteinuria in a screenedvolunteers [1], and reported as a risk factor for cardiovas-cohort of subjects with normal kidney function.

Methods. A total of 5403 subjects (3403 men and 2000 women) cular or total mortality [2, 3]. Microalbuminuria is awho participated in the 1997 and 1999 health screening exami- predictor of nephropathy in diabetic patients [4, 5] andnations in Okinawa Japan, and who were normal renal function is associated with renal functional abnormalities in non-(serum creatinine �1.2 mg/dL in men, �1.0 mg/dL in women)

diabetic subjects [6]. Smoking was reported as an associ-and negative proteinuria by dipstick examination in 1997 wereated factor with microalbuminuria, independently ofeligible for study. Logistic analysis was used to examine the

relation between the baseline state of smoking or obesity in blood pressure, in a cross sectional study including hyper-1997, and the development of proteinuria in 1999, adjusted for tensive and non-hypertensive patients [7], and in type 1age, sex, and other confounding factors. [8] and type 2 diabetes mellitus [9–11]. Furthermore,

Results. Proteinuria developed in 5.8% of participants (6.7% Orth et al’s retrospective study showed that smokingin men, 4.4% in women; dipstick score, 1� in 277, 2� in 37, andincreased the risk of ESRD in men diagnosed with IgA�3� in 4 participants). The incidence of proteinuria was posi-nephropathy or autosomal dominant polycystic kidneytively associated with the number of cigarettes smoked per day

(P � 0.04), and a body mass index (P � 0.0001) at baseline. disease [12]. A large number of reports on the effectsAnalysis showed that the relative risk (95% confidence inter- of smoking on the kidney are available; however, studiesval) of developing proteinuria was 1.32 (1.00 to 1.74), P � 0.04 concerning the impact of smoking on renal function infor cigarette smoking, 1.45 (1.13 to 1.86), P � 0.002 for obesity,

subjects without renal disease are scarce [13, 14].1.56 (1.19 to 2.06), P � 0.001 for hypertension, and 2.27 (1.55In hypertensive and non-hypertensive patients diag-to 3.32), P � 0.0001 for diabetes mellitus. Stratified with men

and women, the relative risk was 1.28 (0.96 to 1.72), P � 0.09 nosed with type 2 diabetes mellitus, obesity was associ-for smoking, and 1.60 (1.19 to 2.14), P � 0.001 for obesity in ated with microalbuminuria [10, 11, 15], and urinarymen; the relative risk was 1.30 (0.44 to 3.80), P � 0.62 for albumin excretion has been positively and significantlysmoking, and 1.04 (0.63 to 1.72), P � 0.87 for obesity in women.

correlated with body weight [16]. Compared with the con-Conclusions. Hypertension and diabetes mellitus were supe-trol group, urinary albumin excretion was significantlyrior to smoking and obesity in predicting the development of

proteinuria in all subjects. Stratified with men and women, higher in the obese patients [17]. In large cross-sectionalobesity was a significant risk factor for the development of pro- health screening studies of middle-aged men and women

known to be non-diabetic, non-hypertensive, non-lipid-emic, and non-proteinuric, obesity correlated with uri-Key words: cigarettes, tobacco, end-stage renal disease, overweight,nary albumin excretion [18]. Furthermore, a significanthypertension, diabetes mellitus, cardiovascular disease, blood pressure,

renal failure. correlation has been observed between body weight lossand a decrease in proteinuria [19]. However, a largeReceived for publication June 18, 2001follow-up study has not been available.and in revised form March 25, 2002

Accepted for publication March 28, 2002 Therefore, we examined the effects of smoking andobesity on the development of proteinuria in a large co- 2002 by the International Society of Nephrology

956

Tozawa et al: Smoking, obesity, risk of proteinuria 957

hort of subjects with normal renal function and without or the current use of antihypertensive medication. Obe-sity was defined as body mass index (BMI) �25 kg/m2,proteinuria. The established risk of hypertension and

diabetes mellitus for renal failure and ESRD [20, 21] calculated as body weight (kg) divided by the square ofthe height (m2) [25]. Diabetes mellitus was diagnosed ifwas adjusted for in the statistical analysis of the data, so

that the true effect of smoking and obesity on proteinuria fasting blood glucose level was �126 mg/dL (7 mmol/L)[26], HbA1c was �7.0%, or the subject was under dia-could be observed independently of these variables.betes mellitus therapy. Hypercholesterolemia was de-fined as total cholesterol �220 mg/dL (5.69 mmol/L)

METHODS[27]. Hypertriglyceridemia was defined as triglycerides

The characteristics of this ongoing screening program �150 mg/dL (1.7 mmol/L) [28]. Anemia was defined asin which subjects in the present study participated have blood hemoglobin concentration less than mean � 2 SDbeen reported previously [22, 23]. The screening is one- of the subjects (�14.0 g/dL in men, and �11.0 g/dL inday clinic program run by the Okinawa General Health women). Hyperuricemia was defined as serum uric acidMaintenance Association, and involves a thorough phys- �7.0 mg/dL (420 �mol/L) [29].ical examination for about 10,000 adults who presentlyparticipate each year. Among 9914 subjects screened in Statistical methods1997, 8419 had a normal renal function [serum creatinine The unpaired t test or 2 test was used to compare(SCr) �1.2 mg/dL in men, �1.0 mg/dL in women] and values or ratios between groups. Trends of frequency werenegative proteinuria by dipstick examination. Of these tested by multiple regression analysis or analysis of vari-normal renal function subjects, 5403 subjects partici- ance (ANOVA). Difference of distribution for frequencypated in the repeat 1999 program, but 3016 subjects between dropped and included subjects also was testeddid not. Finally, a total of 5403 subjects (3403 men, 2000 by the Wilcoxon rank sum test. Relative risk of protein-women) were included in the analysis for the present uria was estimated with the use of logistic regressionstudy. This study was conducted in accordance with the models (StatView 5.0; SAS Institute, Cary, NC, USA).principles set out in the Declaration of Helsinki 1975, These particular variables for the multivariable logisticas revised in 1993. The study design was approved by regression model were selected because they were thoughtthe ethical committee of the Association. Only the field to be related to cardiovascular risk, mortality risk, renaldata set that excluded information such as name and ad- risk and renal function. Logistic analysis was used to finddress was used for this study. After the nature of screen- the predictor of developing proteinuria. Table 2 lists theing was explained, all participants signed the question- variables included in the multivariate analysis. We didnaire. not use a stepwise selection method. P � 0.05 was consid-

The following parameters were noted in the self- ered significant.administered, standardized questionnaire: family historyof hypertension, presence of diabetes mellitus, current

RESULTScigarette smoking habit (0, 1 to 10, 11 to 20, and �21cigarettes per day), level and incidence of alcohol con- Basal characteristics in 1997 of subjects who dropped

out and those who were included in analysis are shownsumption, and the amount of physical exercise under-taken. To corroborate this information these subjects were in Table 1. Serum creatinine and the prevalence of hy-

pertriglyceridemia, hyperuricemia, and current drinkerinterviewed by a physician. Dipstick urinalysis (Amesdipstick) was performed on fresh, mid-stream urine col- were significantly higher in subjects who were included

in analysis than in those who dropped out. Age and thelected in the morning. The results of the urine test wereinterpreted as (�), (�/�), (1�), (2�), (3�) and (4�). prevalence of diabetes mellitus were significantly lower

in subjects who were included in analysis than in thoseHeight, body weight, blood pressure, blood hemoglobin(Hb), fasting blood glucose, HbA1c, serum creatinine, who dropped out. Because the number of subjects was

relatively large, the prevalence of some risk factors wasuric acid, total cholesterol, and triglycerides were mea-sured in all subjects. After 15 minutes of sitting, blood statistically different between the two groups (Table 1).

However, no difference was found in the prevalence ofpressure was measured twice by a trained nurse using astandard sphygmomanometer. The lower value of two obesity and smoking between the two groups. Further-

more, the entire trends of the prevalence of risk fac-blood pressure measurements was used for this study.Normal renal function was defined as SCr �1.2 mg/dL tors between subjects who dropped out and those who

were included in analysis were comparable (P � 0.83 byin men and �1.0 mg/dL in women. Negative proteinuriawas defined as a dipstick proteinuria (�) and (�/�), ANOVA). For prevalence of risk factors, there was no

observed difference (P � 0.40) between the droppedand a positive proteinuria was defined as (1�) or more.Hypertension was defined as systolic blood pressure and included subjects by Wilcoxon rank sum test.

In 1999, proteinuria was recorded in 318 of the subjects�140 mm Hg, diastolic blood pressure �90 mm Hg [24],

Tozawa et al: Smoking, obesity, risk of proteinuria958

Table 1. Baseline characteristics of subjects who were included in analysis and subjects with normal renal function andnegative proteinuria in 1997 who dropped out as well as baseline characteristics of those included in the analysis

divided by whether or not they developed proteinuria in 1999

Subjects (N�8419) Proteinuria in 1999 (N�5403)

Dropped out Included in analysis No YesCharacteristics (N�3016) (N�5403) (N�5085) (N�318)

Mean ageSD years 4911 489a 489 48 9Sex % male 52 62a 62 72b

Mean serum creatinineSD mg/dL 0.940.15 0.970.15a 0.97 0.15 0.980.14Obesity % 33 34 33 47b

Hypertension % 23 21 21 31b

Hypercholesterolemia % 34 33 33 36Hypertriglyceridemia % 27 31a 31 42b

Diabetes mellitus % 6.6 5.1a 4.7 12b

Hyperuricemia % 16 18a 18 22Anemia % 6.8 6.3 6.2 8.1Current drinker % 59 65a 65 72b

Current smoker % 27 28 28 36b

Habitual exerciser % 35 36 37 35

Definitions of the clinical characteristics are described in the text. The trends of prevalence of risk factors between subjects dropped out and included in analysiswere comparable (P � 0.83 by ANOVA).

a P � 0.05 vs. subjects dropped outb P � 0.05 vs. no proteinuria

(5.8% total; 6.7% in men, 4.4% in women, P � 0.001 vs. was positive in all subjects (P � 0.0001). The incidenceof proteinuria was significantly associated with BMI formen): 1� (N � 277), 2� (N � 37), �3� (N � 4). Base-

line characteristics of 5403 subjects divided by the pres- men (P � 0.0001), but not in women (P � 0.4) (Fig. 2).Table 2 presents the results of multivariate logisticence or absence of proteinuria in 1999 also are shown

in Table 1. Subjects who developed proteinuria were analysis of predictors for developing proteinuria in sub-jects with normal renal function and negative proteinuriamore likely to be male, obese, and hypertensive. They

were also more likely to have hypertriglyceridemia, dia- (N � 5403). The dependent variable was positive pro-teinuria in 1999. Current smoking, obesity, hypertensionbetes mellitus, drinking and smoking habits. Comparing

the prevalence values in men with those in women among and diabetes mellitus were statistically significant afteradjustment for age, sex, hypercholesterolemia, hypertri-5403 subjects, the prevalence of obesity (men vs. women,

39.2 vs. 25.7%), hypertension (23.7 vs. 20.5%), hypertri- glyceridemia, anemia, hyperuricemia, drinking habitsand exercise habits. Hypertension and diabetes mellitusglyceridemia (40.7 vs. 15.7%), diabetes mellitus (6.5 vs.

2.9%), hyperuricemia (28.5 vs. 2.0%), current drinker were superior to smoking and obesity in predicting thedevelopment of proteinuria.(87.7 vs. 26.8%), current smoker (43.0 vs. 4.5%), and

habitual exerciser (39.6 vs. 31.6%) were higher in men A sex difference may be present for the relationshipbetween smoking or obesity and the development ofthan in women (P � 0.0001, respectively).

We did not analyze the relationship between protein- proteinuria (Figs. 1 and 2). Therefore, we also analyzedthe data separately in men (N � 3403, mean age SD,uria and the risk factors in 1999 because complete data

set in 1999 was not available. However, the body mass 47 9 years old) and women (N � 2000, 50 9 yearsold; Table 3). Current smoking was significant in theindex (BMI) in 1999 was obtained. According to the

data, 34% of all subjects were obese in 1997; 31% of all unadjusted model and showed a trend in the adjustedmodel in men, but was not significant in women. Obesitysubjects maintained obese from 1997 to 1999, and 4%

of all subjects newly developed obesity in 1999. and diabetes mellitus were significant predictors for de-veloping proteinuria only in men but not in women.The correlation between incidence of newly developed

proteinuria and the baseline number of cigarettes smoked There was a trend for hypertension in men (P � 0.05in the adjusted model). Hypertension was a significantper day was significantly positive in all subjects (P �

0.04), adjusted for age and sex (Fig. 1). Since the number predictor for the development of proteinuria in women(P � 0.002 in adjusted model).of women who smoked in the sample population was

small, no correlation was evident in women. Figure 2 The data were analyzed after excluding the hyperten-sive and diabetic patients in 1997 from the study group.shows the incidence of proteinuria in 1999 according to

the baseline BMI in 1997 divided by the WHO classifica- In all subjects (2465 men, 45 9 years old; 1589 women,48 9 years old; mean age SD), the relative risk oftion of obesity [25]. After being adjusted for age and

sex, the correlation of incidence of proteinuria to BMI developing proteinuria was 1.33 (0.96 to 1.85; P � 0.08)

Tozawa et al: Smoking, obesity, risk of proteinuria 959

Fig. 1. Incidence of proteinuria according to the number of cigarettesFig. 2. Incidence of proteinuria according the baseline body mass indexsmoked per day. Trends were significant for all subjects (�; P � 0.04)(BMI). Trends were significant for all subjects (�; P � 0.0001) when ad-when adjusted for age and sex, but not significant when adjusted forjusted for age and sex, and when adjusted for age in men ( ; P � 0.0001),age in men ( ; P � 0.07) or in women (�; P � 0.9).but not significant when adjusted for age in women (�; P � 0.4).

Table 2. Results of the logistic analysis of developing proteinuria in 5403 subjects with normal renal function and negative proteinuria

Not adjusted Adjusteda

Predicting factors � RR (95% CI) P � RR (95% CI) P

Current smoking (vs. non-smoker) 0.38 1.47 (1.15–1.86) 0.001 0.28 1.32 (1.00–1.74) 0.04Obesity (vs. non-obesity) 0.43 1.54 (1.22–1.95) 0.0003 0.37 1.45 (1.13–1.86) 0.002Hypertension (vs. normotension) 0.41 1.51 (1.17–1.95) 0.001 0.45 1.56 (1.19–2.06) 0.001Diabetes mellitus (vs. non-diabetes) 0.82 2.28 (1.58–3.31) �0.0001 0.82 2.27 (1.55–3.32) �0.0001

Abbreviations are: RR, relative risk; CI, confidence interval.a Adjusted for other predicting factors in the Table and other confounding variables such as age, sex, hypercholesterolemia, hypertriglyceridemia, anemia,

hyperuricemia, drinking habits, and exercise habit. The definitions of these factors are described in the text.

for smoking, and 1.51 (1.11 to 2.07; P � 0.008) for obesity glucose or HbA1c measurements, respectively. In all sub-jects (2241 men, 45 9 years old; 1492 women, 48 9after an adjustment for age, sex, hypercholesterolemia,

hypertriglyceridemia, anemia, hyperuricemia, drinking years old; mean age SD), the relative risk of devel-oping proteinuria was 1.40 (0.95 to 2.06; P � 0.08) forhabits and exercise habits. When men and women were

stratified for smoking, the adjusted relative risk of devel- smoking, and 1.45 (1.02 to 2.06; P � 0.03) for obesityafter an adjustment with age, sex, hypercholesterolemia,oping proteinuria was 1.25 (0.87 to 1.81; P � 0.22) in

men and 1.50 (0.50 to 4.41; P � 0.46) in women. For hypertriglyceridemia, anemia, hyperuricemia, drinkinghabits and exercise habits. When men and women wereobesity the relative risk was 1.74 (1.20 to 2.51; P � 0.003)

in men and 0.98 (0.51 to 1.88; P � 0.95) in women. stratified for smoking, the adjusted relative risk of devel-oping proteinuria was 1.36 (0.89 to 2.06; P � 0.14) inThese data showed obesity was a significant predictor

for developing proteinuria in non-hypertensive and non- men and 1.59 (0.53 to 4.74; P � 0.40) in women; and forobesity the adjusted relative risk was 1.74 (1.15 to 2.64;diabetic men.

We also analyzed the data after excluding the hyper- P � 0.008) in men and 0.81 (0.38 to 1.73; P � 0.60) inwomen. These data suggested that obesity was a signifi-tensive and diabetic patients in 1997 or 1999 from the

study group. However, since data for hypertensive or cant predictor for developing proteinuria in non-hyper-tensive and non-diabetic men.diabetic therapy in 1999 were not obtained, the diagnosis

of hypertension or diabetes mellitus in 1999 was deter- To assess the probability of developing proteinuria ac-cording to risk factor-clustering, two sets of logistic mod-mined only by the blood pressure values and fasting blood

Tozawa et al: Smoking, obesity, risk of proteinuria960

Table 4. Odds ratio (95% confidence interval) for developingproteinuria based on the number of risk factors: smoking,

obesity, hypertension, and diabetes mellitusa

Odds ratio (95% confidence interval)

Number of risk factors Not adjusted Adjustedb

None (N�2018) 1.00 1.00One (N�2130) 1.79 (1.33–2.41) 1.67 (1.22–2.29)Two (N�995) 2.39 (1.72–3.33) 2.18 (1.51–3.14)Three (N�214) 4.12 (2.59–6.55) 3.82 (2.31–6.29)Four (N�21) 13.71 (5.36–35.02) 12.21 (4.67–31.89)

a Patients whose smoking habits were not known in 1997 (N � 25) were excludedb Adjusted for age, sex, hypercholesterolemia, hypertriglyceridemia, anemia,

hyperuricemia, drinking habit, and exercise habit. The definitions of these factorsare described in the text.

els were used (Table 4). Each assessed the single andcumulative number of risk factors from significant pre-dictors of proteinuria: smoking, obesity, hypertension,and diabetes mellitus. Adjustments were made for age,sex, hypercholesterolemia, hypertriglyceridemia, ane-mia, hyperuricemia, alcohol-drinking status, and physicalexercise. The magnitude of the estimates in the two setsof models showed that an increasing number of riskfactors indicated a higher risk of proteinuria.

DISCUSSION

Smoking is a significant risk factor of proteinuria invarious clinical settings such as hypertension, non-hyper-tension [7], type 1 diabetes [8], type 2 diabetes [9–11],and screened men [3]. Halimi et al’s study of 28,409volunteers reported in that the unadjusted odds ratio(95% confidence interval) of a current or former smokerwas 2.15 (1.56 to 3.00) for a dipstick score of 1� orhigher proteinuria [14]. From the Multiple Risk FactorIntervention Trial (MRFIT) cohort, Grimm et al re-ported that smoking predicted the incidence of 1� orhigher proteinuria through 72 months in 12,866 men;the odds ratio was 1.28 (1.14 to 1.43) with multivariateadjustments [3]. In our present study, current cigarettesmoking predicts the development of proteinuria in allsubjects independent of hypertension and diabetes melli-tus. After a stratification of men and women, there wasonly a trend in men (N � 3403) and no significance inwomen (N � 2000). The relatively small number of sub-jects may limit the statistical power, however.

A large body of literature has remarked on the po-tential adverse effects of cigarette smoking on health[12, 13]. Infusion of nicotine into the renal artery in-creases glomerular filtration rates (GFR) [30]. In a largecross-sectional study (N � 7476), the percentage of sub-jects with an elevated GFR was significantly larger incurrent smokers (5.4%) than non-smokers (3.0%) [31].An increase in GFR induced by smoking could generate

Tab

le3.

Res

ults

ofth

elo

gist

ican

alys

isof

deve

lopi

ngpr

otei

nuri

ain

5403

subj

ects

wit

hno

rmal

rena

lfu

ncti

onan

dne

gati

vepr

otei

nuri

a

Not

adju

sted

Adj

uste

da

Men

Wom

enM

enW

omen

Pre

dici

tng

fact

or�

RR

(95%

CI)

p�

RR

(95%

CI)

p�

RR

(95%

CI)

p�

RR

(95%

CI)

p

Cur

rent

smok

ing

(vs.

non

smok

er)

0.30

1.35

(1.0

3–1.

78)

0.02

0.12

1.12

(0.4

0–3.

16)

0.81

0.25

1.28

(0.9

6–1.

72)

0.09

0.26

1.30

(0.4

4–3.

80)

0.62

Obe

sity

(vs.

non

obes

ity)

0.53

1.70

(1.2

9–2.

24)

0.00

010.

041.

04(0

.64–

1.69

)0.

860.

471.

60(1

.19–

2.14

)0.

001

0.04

1.04

(0.6

3–1.

72)

0.87

Hyp

erte

nsio

n(v

s.no

rmot

ensi

on)

0.27

1.31

(0.9

7–1.

78)

0.07

0.78

2.19

(1.3

6–3.

53)

0.00

10.

311.

36(0

.98–

1.89

)0.

050.

792.

21(1

.32–

3.69

)0.

002

Dia

bete

sm

ellit

us(v

s.no

ndi

abet

es)

0.91

2.50

(1.6

8–3.

73)

�0.

0001

0.20

1.23

(0.4

2–3.

56)

0.69

0.97

2.65

(1.7

6–3.

99)

�0.

0001

�0.

160.

85(0

.25–

2.88

)0.

64

The

data

wer

ean

alyz

edse

para

tely

inm

enan

dw

omen

.Abb

revi

atio

nsar

e:R

R,r

elat

ive

risk

;CI,

conf

iden

cein

terv

al.

aA

djus

ted

wit

hot

her

pred

icti

ngfa

ctor

sin

the

Tab

lean

dot

her

conf

ound

ing

vari

able

ssu

chas

sex,

hype

rcho

lest

erol

emia

,hy

pert

rigl

ycer

idem

ia,

anem

ia,

hype

ruri

cem

ia,

drin

king

habi

ts,

and

exer

cise

habi

t.T

hede

fini

tion

sof

thes

efa

ctor

sar

ede

scri

bed

inth

ete

xt.

hyperfiltration, a risk factor for progression to chronic

Tozawa et al: Smoking, obesity, risk of proteinuria 961

Table 5. Frequency of proteinuria examined by dipstick test in relatively large studies

References Area surveyed Year of survey N (men/women) Age Frequency of proteinuria

Culleton et al [2] USA (Framingham) 1977 2586 (1045/1541) 68 Men (6.3%), women (3.4%)a

Grimm et al [3] USA (MRFIT) 1973–1980 12,866 (men) 35–57 3.3–6.7%,a 2.4%/yearb

Halimi et al [14] France 1998 28,409 20–69 0.7%a

Von Bonsdorff et al [40] Finland 1976 36,147 (men) 20 0.4%a

Tozawa et al [present study] Japan (Okinawa) 1997 9908 18–89 6.4%a

1997–1999 5403 (3403/2000) 48 Men (3.3%/year), women (2.2%/year)b

Age represents mean or range (years).a Prevalenceb Incidence

renal disease [32]. Furthermore, secretion of vasopressin abetes or hypertension may be greater than obesity andsmoking. In the present study, the risk of developingis increased during smoking, and vasopressin increases

GFR inappropriately via changing the activity of tubu- proteinuria was also increased by the combination of fourrisk factors: smoking, obesity, hypertension, and diabetesloglomerular feedback [33].

Proteinuria is relatively common in obese patients. mellitus. Ribstein, du Cailar and Mimran also showedthat obesity magnified the effects of hypertension on al-Dornfeld reviewed 1000 obese subjects and revealed 410

(41.0%) cases of proteinuria with normal urine sedi- buminuria [35]. Together these results suggest that mul-tiple risk factors increase the risk of developing pro-ments [34]. Body mass index was strongly correlated with

microalbuminuria in a cross-sectional population-based teinuria.Frequency of proteinuria examined by the dipstickstudy [18]. Praga et al reported that hypocaloric diets

decreased body weight as well as proteinuria in nine test is different in surveyed areas worldwide (Table 5).The prevalence of proteinuria in the USA [2, 3] andobese patients [19]. Proteinuria in obesity may be benign;

however, it was reported that out of 1000 obese patients Japan (the present study) were higher than in France[14] and Finland [40]. However, the frequency of protein-53 (5.3%) had an abnormal urinary sediment [34]. A

larger study to determine the effect of weight reduction uria was comparable between USA and Japan, or Franceand Finland. Factors including race, environmental back-on proteinuria should be carried out.

The mechanisms by which obesity induce renal injury grounds, age, and rate of men in the studied cohorts maycontribute to the difference. The prevalence of diabetesare poorly understood. Obesity is associated with renal

hyperfiltration and hyperperfusion [35]. Kidneys from mellitus and hypertension also could account for the dis-crepancies among these studies; for example, the preva-obese subjects exhibit focal glomerulosclerosis and other

histologic changes similar to those observed in diabetic lence of hypertension was 8.9% in France [14] versus21% in Japan (present study).nephropathy [36]. A recent experimental study in dogs

showed that a high-fat diet caused obesity, glomerular Potential limitations should be considered in the inter-pretation of this study. First, the duration of proteinuriahyperfiltration and glomerular structural change associ-

ated with a trend for an elevated glomerular transform- that had persisted in each subject before 1997 was un-known. This may be important since transient protein-ing growth factor (TGF)-�1 expression [37].

In the present study, the incidence of developing pro- uria with a resolution on subsequent examination wasreported to be 4% in men and 7% in women [41]. Second,teinuria was greater in men than in women; this is proba-

bly due to the prevalence of cigarette smoking, obesity, it might be uncertain that a single dipstick urine testcould reflect organ damage and thus be a reliable prog-hypertension and diabetes mellitus in the male popula-

tion as a whole. This gender difference is further high- nostic test for the progression of renal disease. How-ever, Davidson and Smiley reported that a 1� or higherlighted as men are more likely to progress to renal disease

[38], and it has been shown worldwide that a higher dipstick-positive proteinuria was highly predictive ofabnormal urinary albumin/creatinine ratios [42]. Withnumber of men are accepted for dialysis than women.

Animal studies have shown that sex-related hormones microalbuminuria, a chance dipstick measurement ofproteinuria predicted ESRD; a dose-dependent relation-were involved when explaining gender differences in the

incidence of proteinuria and glomerulosclerosis [39]. ship was observed between the degree of dipstick pro-teinuria and a risk of ESRD [1]. Furthermore, anotherIn all subjects diabetes mellitus is followed by hyper-

tension, obesity and smoking as the most important risk study found that casual dipstick proteinuria was a sig-nificant risk factor for all-cause mortality [2].factors for developing proteinuria (Table 4). Hyperten-

sion in women or diabetes mellitus men was the strongest In summary, hypertension in women, and diabetesmellitus in men were significant risk factors for devel-predictor of the forthcoming proteinuria in each group

(Table 3). It should be emphasized that influence of di- oping proteinuria. Hypertension in men showed a trend

Tozawa et al: Smoking, obesity, risk of proteinuria962

18. Metcalf P, Baker J, Scott A, et al: Albuminuria in people atto increase the risk of developing proteinuria. Obesityleast 40 years old: Effect of obesity, hypertension, and hyperlipid-was a significant risk factor for the development of pro- emia. Clin Chem 38:1802–1808, 1992

teinuria independently of both hypertension and diabe- 19. Praga M, Hernandez E, Andres A, et al: Effects of body-weightloss and captopril treatment on proteinuria associated with obesity.tes mellitus in men. The risk of developing proteinuriaNephron 70:35–41, 1995also tended to be increased with cigarette smoking in 20. Klein R, Klein BE, Moss SE, et al: The 10-year incidence of renal

men. Smoking and obesity in women were not significant insufficiency in people with type 1 diabetes. Diabetes Care 22:743–751, 1999in this data set.

21. Ritz E: Nephropathy in type 2 diabetes. J Intern Med 245:111–126, 1999ACKNOWLEDGMENTS 22. Tozawa M, Oshiro S, Iseki C, et al: Multiple risk factor clus-tering of hypertension in a screened cohort. J Hypertens 18:1379–We thank Mr. M. Itokazu and Mr. K. Shiroma of the Okinawa1385, 2000General Health Maintenance Association.

23. Tozawa M, Oshiro S, Iseki C, et al: Family history of hyperten-sion and blood pressure in a screened cohort. Hypertens Res 24:93–Reprint requests to Masahiko Tozawa, M.D., Third Department of98, 2001Internal Medicine, University of The Ryukyus, 207 Uehara, Nishihara-

24. Guidelines Subcommittee of the Japanese Society of Hyperten-cho, Okinawa 903-0215, Japan.sion: Japanese Society of Hypertension Guidelines for the Manage-E-mail: tozawa@med.u-ryukyu.ac.jpment of Hypertension (JSH 2000). [Japanese] Tokyo, Japanese So-ciety of Hypertension, 2000REFERENCES 25. Global prevalence and secular trends in obesity, in Obesity. Pre-venting and Managing the Global Epidemic—Report of a WHO1. Iseki K, Iseki C, Ikemiya Y, Fukiyama K: Risk of developing end-Consultation on Obesity (vol 3). Geneva, World Health Organiza-stage renal disease in a cohort of mass screening. Kidney Int 49:tion, 1997800–805, 1996

26. The Expert Committee on the Diagnosis and Classification2. Culleton BF, Larson MG, Parfrey PS, et al: Proteinuria as aof Diabetes Mellitus: Report of the expert committee on therisk factor for cardiovascular disease and mortality in older people:diagnosis and classification of diabetes mellitus. Diabetes Care 20:A prospective study. Am J Med 109:1–8, 20001183–1197, 19973. Grimm RH Jr, Svendsen KH, Kasiske B, et al: Proteinuria is a

27. Teramoto T: Guideline for diagnosis and management of hyperlip-risk factor for mortality over 10 years of follow-up. Multiple Riskidemia for prevention of atherosclerosis in Japan. Nippon RinshoFactor Intervention Trial. Kidney Int 52(Suppl 63):S10–S14, 199757:2651–2656, 19994. Viberti GC, Hill RD, Jarrett RJ, et al: Microalbuminuria as a

28. Castelli W: Epidemiology of triglycerides: A view from Framing-predictor of clinical nephropathy in insulin-dependent diabetesmellitus. Lancet 1:1430–1432, 1982 ham. Am J Cardiol 70:3H–9H, 1992

5. Mogensen CE: Microalbuminuria predicts clinical proteinuria and 29. Wortmann R: Gout and other disorders of purine metabolism, inearly mortality in maturity-onset diabetes. N Engl J Med 310:356– Harrison’s Principles of Internal Medicine (14th ed), edited by360, 1984 Fauci AS, Braunwald E, Isselbacher KJ, et al: New York,

6. Pinto-Sietsma SJ, Janssen WM, Hillege HL, et al: Urinary albu- McGraw-Hill, 1998, pp 2158–2166min excretion is associated with renal functional abnormalities in 30. Pawlik WW, Jacobson ED, Banks RO: Actions of nicotine ona nondiabetic population. J Am Soc Nephrol 11:1882–1888, 2000 renal function in dogs. Proc Soc Exp Biol Med 178:585–590, 1985

7. Horner D, Fliser D, Klimm HP, Ritz E: Albuminuria in normo- 31. Pinto-Sietsma SJ, Mulder J, Janssen WM, et al: Smoking is relatedtensive and hypertensive individuals attending offices of general to albuminuria and abnormal renal function in nondiabetic persons.practitioners. J Hypertens 14:655–660, 1996 Ann Intern Med 133:585–591, 2000

8. Stegmayr B, Lithner F: Tobacco and end stage diabetic nephropa- 32. Neuringer JR, Brenner BM: Hemodynamic theory of progressivethy. BMJ 295:581–582, 1987 renal disease: A 10-year update in brief review. Am J Kidney Dis

9. Olivarius NDF, Andreasen AH, Keiding N, Mogensen CE: 22:98–104, 1993Epidemiology of renal involvement in newly-diagnosed middle- 33. Bankir L, Bouby N: Vasopressin and urinary concentration: Addi-aged and elderly diabetic patients. Cross-sectional data from the tional risk factors in the progression of chronic renal failure. Ampopulation-based study “Diabetes Care in General Practice,” Den- J Kidney Dis 17:20–26, 1991mark. Diabetologia 36:1007–1016, 1993 34. Dornfeld L: Obesity, in Textbook of Nephrology (2nd ed), edited

10. Savage S, Nagel NJ, Estacio RO, et al: Clinical factors associated by Glassock RJ, Massry SG, Baltimore, Williams and Willkins,with urinary albumin excretion in type II diabetes. Am J Kidney 1989, pp 989–995Dis 25:836–844, 1995 35. Ribstein J, du Cailar G, Mimran A: Combined renal effects of

11. Esmatjes E, Castell C, Gonzalez T, et al: Epidemiology of renal overweight and hypertension. Hypertension 26:610–615, 1995involvement in type II diabetics (NIDDM) in Catalonia. The Cata- 36. Kasiske BL, Crosson JT: Renal disease in patients with massivelan Diabetic Nephropathy Study Group. Diabetes Res Clin Pract obesity. Arch Intern Med 146:1105–1109, 199632:157–163, 199637. Henegar JR, Bigler SA, Henegar LK: et al Functional and struc-12. Orth SR, Stockmann A, Conradt C, et al: Smoking as a risk

tural changes in the kidney in the early stages of obesity. J Am Socfactor for end-stage renal failure in men with primary renal disease.Nephrol 12:1211–1217, 2001Kidney Int 54:926–931, 1998

38. Gretz N, Zeier M, Geberth S, et al: Is gender a determinant13. Orth SR, Ritz E, Schrier RW: The renal risks of smoking. Kidneyfor evolution of renal disease? A study in autosomal dominantInt 51:1669–1677, 1997polycystic kidney disease. Am J Kidney Dis 13:178–183, 198914. Halimi JM, Giraudeau B, Vol S, et al: Effects of current smoking

39. Sakemi T, Toyoshima H, Morito F: Testosterone eliminates theand smoking discontinuation on renal function and proteinuria inattenuating effect of castration on the progressive glomerularthe general population. Kidney Int 58:1285–1292, 2000injury in hypercholesterolemic male Imai rats. Nephron 67:469–15. Spangler JG, Konen JC: Hypertension, hyperlipidemia, and ab-476, 1994dominal obesity and the development of microalbuminuria in pa-

40. von Bonsdorff M, Koskenvuo K, Salmi HA, Pasternack A:tients with non-insulin-dependent diabetes mellitus. J Am BoardPrevalence and causes of proteinuria in 20-year-old Finnish men.Fam Pract 9:1–6, 1996Scand J Urol Nephrol 15:285–290, 198116. Reid M, Bennett F, Wilks R, Forrester T: Microalbuminuria,

41. Robinson RR: Isolated proteinuria in asymptomatic patients. Kid-renal function and waist:hip ratio in black hypertensive Jamaicans.ney Int 18:395–406, 1980J Hum Hypertens 12:221–227, 1998

42. Davidson MB, Smiley JF: Relationship between dipstick positive17. Valensi P, Assayag M, Busby M, et al: Microalbuminuria in obeseproteinuria and albumin:creatinine ratios. J Diabetes Complica-patients with or without hypertension. Int J Obes Relat Metab

Disord 20:574–579, 1996 tions 13:52–55, 1999