Transcript Slide 1

Gender-specific protective effect
of hemoglobin on arsenicinduced skin lesions
David C. Christiani,MD,MPH,MS
Dhaka, Bangladesh
January 4, 2006
Harvard-DCH Collaborators
• Carrie V. Breton, Andres Houseman, Molly
Kile, Quazi Quamruzzaman, Mahmuder
Rahman, Golam Mahiuddin, David C.
Christiani.
• Dhaka Community Hospital, Dhaka,
Bangladesh and Harvard School of Public
Health, Boston, Mass., USA
Background
• Chronic arsenic poisoning is a major health concern
in Bangladesh and in other regions of the world.
• Mechanisms of human toxicity remain elusive,
especially dermatologic outcomes.
• In vitro and in vivo studies show that inorganic
arsenic binds to hemoglobin and can change cell
morphology, heme metabolism, and hemoglobin
levels.
Background
• Acute exposure to arsernite has been shown to
cause anemia, leukopenia, and throbocytopenia
from bone marrow suppression.
• Arsine gas can cause hemolytic anemia.
• Arsenic-induced skin lesions, cancerous
lesions, have less hemoglobin than do noncancerous lesions.
Background
• Chronic arsenic ingestion alters heme
metabolism and increases excretion of total
porphyrins.
• However, the effect of chronic As exposure on
the hemologic system remains relatively
unexplored, with no direct evidence of As
related anemia.
Hemoglobin
• Hemoglobin (Hgb) is of particular interest
because of its use in the diagnosis of anemia.
• Anemia is defined as a blood Hgb below 12g/dL.
• Bangladesh: National prevalence of anemia has
remained constant at 74% for the past 30 years.
• Iron deficiency due to inadequate intake and low
dietary bioavailability of iron.
Aims
• To assess the relationship between blood Hgb and
skin lesions in a case-control study of 1,800 persons
in Pabna.
• To assess potential interactions between AS, Hgb, and
genetic polymorphisms in genes important for
detoxification.
• To examine direct associations between toenail and
urinary arsenic levels and As in a disease-free,
repeated measures study of 184 adults from 50
families is Pabna.
Methods
• Case-Control Study:2001-2003, 900 case-control
pairs (n=1,800) recruited
• Physician-diagnosed lesions; keratosis, spotted
melanosis, Bowen’s disease, SCC
• Separate study: 248 persons (148 adults) from 50
families (25 in suspected high, 25 in suspected
low exposure regions) for a four time per year
sampling over 2001-2005. “Dosimetry Study”
Data collection and analysis
• Hgb determined by Sahli’s method.
• Toenail clippings- Total Inorganic As analyzed by
ICP-MS.
• Well-water sampling after several minute flush.
• Blood for DNA – PCR for GSTM1, GSTT1, GSTP1.
Statistical Analysis
• Descriptive statistics (Chi square, t-test,Wilcoxon
rank sum), comparing Cases vs Controls; and
characteristics of dosimetry group
• Conditional logistic regression to examine Hgb-skin
lesions while adjusting for co-variation (age, gender,
smoking, betel-nut chewing, etc.).
• For repeated measures, separate regressions (mixed
models) with a random family effect.
Table1. Sociodemographic characteristics of
1800 cases and controls
Cases
Controls
P-value
61.7%
61.8%
0.96*
Mean age (years)
33.5
33.3
0.76+
Mean body mass index
20.0
20.2
0.05+
Physical
Male
Social
<0.0001*
Education
no school (illiterate)
20.5%
15.0%
no school (able to write)
30.4%
24.7%
primary
12.2%
13.0%
Secondary
32.6%
41.4%
college +
4.3%
5.9%
Chew tobacco
14.8%
14.8%
0.96*
Chew betel nuts
28.6%
23.9%
0.03*
Ever smoked
28.6%
31.6%
0.17*
Table1 (cont.) Sociodemographic
characteristics of 1800 cases and controls
Cases
Controls
P-value
Demographic
0.0175*
Hemoglobin categories
Hb <12 g/dl (anemic)
13.7%
10.5%
Hb between12.1 - 14.0
71.0%
70.1%
Hb between 14.1 - 18.0
15.4%
19.4%
Median water arsenic(ug/l)
39
11.4
<0.0001§
Median toenail arsenic
(ug/g)
3.7
1.6
<0.0001§
*
Chi square test
+
Student t- test
§
Wilcoxon Rank Sum test
Table 2. Results from a penalized conditional logistic
regression spline model with the outcome skin lesions.
The estimates for Hb are valid above a level of 12 g/dL
(n=1708).
95% CI
OR
lower
upper
Hb (threshold at 12 g/dl)
0.79
0.69
0.91
BMI
1.01
0.97
1.05
no school (able to write)
1.18
0.80
1.73
primary
0.81
0.50
1.30
Secondary
0.48
0.30
0.74
college +
0.90
0.44
1.85
Smoked cigarettes
0.80
0.53
1.22
Smoked bidi
0.48
0.31
0.74
Chewed betelnuts
1.44
1.00
2.09
Chewed tobacco
0.49
0.31
0.78
Logged toenail arsenic
1.79
1.52
2.10
Education
Table 3 (cont.) Results from a penalized conditional logistic
regression spline model with the outcome skin lesions and
a gender by Hb interaction (n=1708). The estimates for Hb
are valid above a level of 12 g/dL.
95% CI
OR
lower
upper
Hemoglobin (threshold at 12)
1.16
0.92
1.46
BMI
1.01
0.96
1.05
no school (able to write)
1.24
0.83
1.85
no school (primary)
0.84
0.52
1.38
Secondary
0.45
0.28
0.72
college +
0.35
0.17
0.73
Education
Table 3 (cont) Results from a penalized conditional logistic regression spline model
with the outcome skin lesions and a gender by Hb interaction (n=1708). The
estimates for Hb are valid above a level of 12 g/dL.
95% CI
OR
lower
upper
Smoked cigarettes
0.77
0.49
1.19
Smoked biri
0.45
0.28
0.71
Chewed betelnuts
1.50
1.02
2.19
Chewed tobacco
0.48
0.30
0.77
Logged toenail arsenic
1.82
1.54
2.15
Gender-Hb interaction
0.52
0.38
0.70
1.13
1.04
1.23
Spline (logged water arsenic)
Linear
non-linear (p value <.0001)
Table 4. Summary statistics describing the physical
and sociodemographic characteristics of the repeated
measures population (N=184).
Physical
Gender (%)
Male
45.1%
Female
54.9%
Age (years)
35.7
Body mass index
20.8
Table 4. Summary statistics describing the physical
and sociodemographic characteristics of the repeated
measures (“dosimetry”) population (N=184).
Social
Smoking
24.5%
Education
no school (illiterate)
16.9%
no school (able to write)
31.5%
primary
20.1%
Secondary
25.0%
college +
6.5%
Demographic
Mean hemoglobin (g/dL)
Median averaged As toenail (μg/g)
Median total urinary As (ppb)
13.7
0.879
29.8
Table 5. Results of a mixed linear regression model with the
outcome Hb (g/dL).
β coef
SE
P value
Logged average toenail arsenic
-0.107
0.146
0.465
Age
-0.006
0.008
0.490
BMI
0.069
0.033
0.040
Gender (Male)
2.234
0.267
<.0001
no school (able to write)
0.074
0.323
0.820
primary
-0.231
0.367
0.531
Secondary
0.001
0.367
0.998
college +
-1.049
0.524
0.048
-0.157
0.312
0.620
Education
Smoked cigarettes
Conclusions
• Low hemoglobin may exacerbate the
detrimental health effects of chronic arsenic
poisoning.
• While providing clean water remains the
primary solution to the arsenic epidemic,
improving nutrition and reducing iron
deficiency may ameliorate the adverse health
effects of arsenic.
Acknowledgments
• We thank our colleagues, technicians,
laboratory, and administrative staff at Dhaka
Community Hospital and the Pabna
Community Clinic in Bangladesh. We also
acknowledge the academic assistance of Tom
Smith and Paul Catalano, and the technical
expertise of Janna Frelich, Thomas Can Geel,
Ian James, Li Su, Ema Rodrigues and
Meredith Jones.
Grant Support
• This publication was made possible by NIH grants
T32 ES07069, research grants ES011622, ES05947,
and center grant # ES00002. Its contents are solely
the responsibility of the authors and do not represent
the official views of the NIH.