American Journal of Epidemiology
Copyright © 1999 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol. 149, No. 12
Printed in U.S.A.
Neural Tube Defects along the Texas-Mexico Border, 1993-1995
Katherine A. Hendricks,1 J. Scott Simpson,2 and Russell D. Larsen1
In response to a 1991 anencephaly cluster in Cameron County, Texas, a surveillance and neural tube defect
(NTD) recurrence prevention project for NTDs was implemented in the 14 Texas-Mexico border counties. For
1993-1995, NTD-affected pregnancies were identified at all gestational ages through active surveillance of
multiple case-ascertainment sources. There were 87 cases of anencephaly, 96 cases of spina bifida, and 14
cases of encephalocele for respective rates of 6.4, 7.1, and 1.1 per 10,000 live births. Of the 197 NTD casewomen, 93% were Hispanic. The overall, Hispanic, and Anglo NTD rates were, respectively, 14.6,14.9, and 10.6
per 10,000 live births. The NTD rate for El Paso County (9.8 per 10,000), the most northwestern Texas county,
was significantly lower (p = 0.001) than the aggregate rate for the rest of the Texas border (17.1 per 10,000).
The overall Texas border rate was significantly higher (p < 0.001) than a recently estimated rate of 9.3 for
California and minimally higher than a recently adjusted rate of 11.3 for the Metropolitan Atlanta Congenital
Defects Program counties (p = 0.052), both of which now reflect all gestational ages. Of the 197 Texas border
cases, 85% (168 cases) reached a gestational age of >20 weeks. Excluding cases of <20 weeks' gestation in
the rate had a more marked effect on reducing the anencephaly rate (4.9 per 10,000) than the spina bifida rate
(6.7 per 10,000). A country of birth was known for 153 (83%) of the 184 Hispanic case-women: 63% were born
in Mexico; 24%, in Texas; and 11%, elsewhere in the United States. Rates for Mexico-born Hispanic women
(15.1 per 10,000) were significantly higher than rates for United States-born Hispanic women (9.5 per 10,000)
(p = 0.006). Am J Epidemiol 1999; 149:1119-27.
anencephaly; Mexican Americans; neural tube defects; spina bifida cystica
rates by race/ethnicity (15, 16). Few Hispanic data are
available for the United States, but it is generally recognized that Hispanic whites have higher NTD rates than
non-Hispanic whites (17, 18), and blacks have lower
rates than either (15).
Until recently, rates in the United States were often
quoted to be about 6 cases per 10,000 live births (19,
20). Those rates, however, included only cases of >20
weeks' gestational age. Since the frequencies of both
prenatal diagnosis and abortion before the gestational
age of 20 weeks have increased, some authors have
begun to adjust rates to include cases of earlier gestational age (21, 22). Recently adjusted rate estimates
including cases of less than 20 weeks' gestational age
have ranged from 7.2 to 15.6 per 10,000 (23).
Although anencephaly rates based on vital records data
have been reported for Texas Hispanics (24, 25) and
NTD rates have been published for Harris County,
Texas (26), the present study is the first to use active
surveillance and multisource case finding to identify
NTD birth outcomes for all gestational ages along the
entire Texas-Mexico border, a population in which 91
percent of all live births occur among white-Hispanic
women of Mexican ancestry.
The following background information was ascertained. Over a 6-week period in 1991, six anencephalic
In humans, the neural tube forms from approximately
gestational days 16 through 27 and is believed to have
multiple closure sites (1,2). Neural tube defects (NTDs),
the most common of which are anencephaly, spina bifida,
and encephalocele, result from multifactorial disturbances in embryonic neurulation (3). Every year in the
United States, about 2,500 infants are born with NTDs
and an additional 1,500 are aborted following prenatal
diagnosis (4). Numerous risk factors for NTDs have been
identified. Exposure to valproic acid/carbamazepine (5),
methotrexate (6), and aminopterin (7); maternal diabetes
(8,9); low socioeconomic status/poverty (10); and hyperthermia (11) have been shown to increase the risk of
NTDs. Genetic risk factors are believed to be important.
There is current interest in the mutated methylenetetrahydrofolate reductase gene (12, 13) and other folate pathway gene variants (14). There is clear variation in NTD
Received for publication April 8, 1996, and accepted for publication September 2,1998.
Abbreviations: ICD-9, International Classification of Diseases,
Ninth Revision; MACDP, Metropolitan Atlanta Congenital Defects
Program; NTD, neural tube defect.
1
Division of Infectious Disease Epidemiology and Surveillance,
Texas Department of Health, Austin, TX.
2
Bureau of Clinical and Nutrition Services, Texas Department of
Health, Austin, TX.
1119
1120
Hendricks et al.
infants were delivered in Cameron County, the southernmost county in the Texas Rio Grande Valley. An
investigation that used active surveillance and multiplesource case finding revealed one of the highest reported
rates in the United States since the early 1970s (19):
27.1 per 10,000 live births for women conceiving from
1990 through 1991 (Texas Department of Health,
unpublished report, 1992). In 1993, in response to the
high rate and community concerns that the cluster was
due to environmental contamination, the Texas
Department of Health implemented NTD surveillance
in the 14 counties along the Texas-Mexico border. This
paper describes the epidemiology of NTDs for women
in these border counties.
MATERIALS AND METHODS
Case finding
Case-women were defined as women who delivered
a live or stillborn infant or had a spontaneous or
induced abortion with a neural tube defect during 1993
through 1995 and had their residence recorded in the
medical record as one of the 14 border counties.
Included in the case series were craniorachischisis (six
cases, coded in the anencephaly group), encephalocele
(14 cases, various cranial lobes), and meningocele (six
cases). Syndromal NTDs (e.g., Meckel-Gruber syndrome) and those associated with a mechanical anomaly (e.g., amniotic band syndrome) or a known chromosomal anomaly (e.g., trisomy or triploidy) were
excluded. Charts of possible NTD cases and casewomen were abstracted from both hospital and nonhospital sources. For NTD outcomes, the woman's
race, ethnicity, and county of residence were based on
these descriptors as they appeared in the medical chart.
Surveillance involved active case ascertainment
from hospitals, birthing centers, genetics clinics, ultrasound centers, abortion centers, prenatal clinics, and
birth attendants including lay midwives and certified
nurse midwives. Additional data sources included various private and government-funded programs targeted
toward physically challenged children. Clinically
apparent cases were ascertained at all gestational ages
and included spontaneous and induced abortions and
stillbirths. In Texas, fetal deaths of natural causes that
occur at gestational ages of <20 weeks are classified as
spontaneous abortions; those that occur at >20 weeks
are classified as stillbirths and must be reported as a
fetal death. Some voluntarily reported spontaneous
abortions are also recorded as fetal deaths.
The hospital surveillance was based upon the
hospital-surveillance protocol of the Metropolitan
Atlanta Congenital Defects Program (MACDP) at the
Centers for Disease Control and Prevention (27, 28).
Each of the 21 hospitals was given a list of
International Classification of Diseases, Ninth
Revision (ICD-9), codes that identified charts for
review and abstraction. The codes included those for
NTDs (those within the ICD-9 designation 740-741
and 742.0), legally and illegally induced abortions,
central nervous system malformations and anomalies,
congenital anomalies, chromosomal abnormalities,
NTD screening procedure codes, stillbirths, and prenatal diagnostic procedures. Medical records personnel pulled charts with any of the screening codes for
review and abstraction by surveillance teams.
Nonhospital surveillance sources were contacted as
follows: 39 birthing centers, four genetic services
providers, and 74 nonhospital outpatient diagnostic
radiology facilities were contacted in person or by
phone at least monthly; four licensed abortion centers
were contacted in person or by phone bimonthly; prenatal clinics and providers were visited at least
bimonthly; and the 103 documented midwives and the
45 licensed certified nurse midwives were contacted in
person or by phone at least every 3 months. Operative
reports were obtained from outside the study area on
infants referred for neurosurgical procedures. To ensure
complete case finding, obstetrician-gynecologists, family practice physicians, pediatricians, neurologists, and
neurosurgeons were surveyed by mail twice during the
study period. The Bureau of Vital Statistics of the Texas
Department of Health also did yearly computer runs for
selected NTD-specific ICD-9 codes for all birth, death,
and fetal death certificates for residents of the 14 border counties. All survey forms and certificates identified through the mailed survey or the vital records
searches were reviewed by the project epidemiologist.
Information for additional possible cases identified
through this procedure was sent to the NTD field
teams. The teams located the medical records to confirm or rule out the NTD diagnoses.
Abstracted charts were reviewed by a physician and
the project epidemiologist. Cases of anencephaly with
spina bifida were counted with the anencephaly cases.
Cases of encephalocele with spina bifida were counted
with the spina bifida cases. Multilevel spina bifida
lesions were coded according to the highest level of
the lesion; that is, lumbosacral lesions were coded as
lumbar. For encephalocele (all ICD-9 code 742.0), the
code specified the affected lobe of the cerebrum.
Confirmed NTD cases were assigned a six-digit, modified ICD-9 code. For anencephaly and similar anomalies (all ICD-9 code 740), the last three digits of the
six-digit code specified the type of anencephaly, craniorachischisis (contiguous opening of brain and spinal
column), or iniencephaly (open occipital malformation
usually accompanied by retroflexion). For spina bifida
Am J Epidemiol Vol. 149, No. 12, 1999
Neural Tube Defects along the Texas-Mexico Border
(all ICD-9 code 741), the last three digits specified the
hydrocephalus status and whether the lesion was open
or closed (fourth digit), the type of spina bifida (fifth
digit), and the highest level of the lesion (sixth digit).
Data analysis
All NTD rates were calculated per 10,000 live births.
The number and demographic characteristics of live
births were obtained from the Bureau of Vital Statistics
of the Texas Department of Health. Any identified
births that occurred outside the study area to residents
of the study area were included in the rate calculations.
For births, demographic variables were based on vital
records descriptors. Confidence intervals (95 percent)
for rates were calculated using the exact binomial (29).
The chi-square for differences in NTD rates, tests for
seasonality, and chi-square tests for trend and multiple
comparisons were calculated using Computer Programs
for Epidemiologic Analysis (30).
An analysis restricted to women with last menstrual
periods during the 24-month period, October 1, 1992,
to September 30, 1994, was performed to analyze seasonality. Because surveillance of outcomes began
January 1, 1993, and because almost no elective abortions of NTD-affected pregnancies occur in Texas
before 12 weeks, the above-mentioned 24-month period
was the earliest period for which outcome data were
complete. To adjust NTD data for any seasonality in
live births, live birth data for the 24-month period 9
months later, July 1, 1993, to June 30, 1995, were used.
RESULTS
During 1993-1995, there were 135,221 resident
live births for the study area; 91 percent were Hispanic
white; 6 percent, Anglo; and 1 percent, non-Hispanic
black. Eighty-nine percent of the births were to women
from the four most populated counties: Cameron,
1121
Hidalgo, Webb, and El Paso. The 10 less populous
counties—Starr, Zapata, Maverick, Kinney, Val Verde,
Terrell, Brewster, Presidio, Jeff Davis, and
Hudspeth—accounted for only 11 percent of the live
births and 11 percent of the NTD cases and hereafter
are considered in aggregate.
NTD cases and rates are summarized by type for
1993-1995 in table 1. The NTD rate was 14.6 per
10,000 live births for the 14-county border region (66
cases in 1993, 68 in 1994, and 63 in 1995). The rate in
the most northwestern county, El Paso County, 9.8 per
10,000 live births, was significantly lower (p = 0.001)
than the aggregate rate (17.1 per 10,000) for the rest of
the border. The El Paso rate also was significantly
lower (p < 0.010) than the individual rates for Webb,
Cameron, and Hidalgo counties of 18.4, 17.4, and
16.2, respectively.
Ninety-three percent (184 cases) of the 197 cases
were Hispanic. Ten cases were Anglo; two, nonHispanic black; and one, Egyptian. Hispanic NTD
rates (14.9 per 10,000) were higher than Anglo NTD
rates (10.6 per 10,000); however, this difference was
not statistically significant (p = 0.29). A country of
birth was known for 153 (83 percent) of the 184
Hispanic case-women; of these, 63 percent were born
in Mexico; 24 percent, in Texas; and 11 percent, elsewhere in the United States. Rates for Mexico-born
Hispanic women (15.1 per 10,000) were significantly
higher than rates for United States-born Hispanic
women (9.5 per 10,000) (p = 0.006). The overall
Hispanic, Mexico-born Hispanic, and United Statesborn Hispanic rates for El Paso were all significantly
lower (p < 0.01) than the overall Hispanic, Mexicoborn Hispanic, and United States-born Hispanic rates
for the other 13 counties considered in aggregate.
When all gestational ages were included for the 3year period, an approximately equal number of anencephaly (n = 87) and spina bifida {n = 96) cases
occurred, with corresponding anencephaly and spina
TABLE 1. Cases by type* and neural tube defect (NTD) ratesf by Texas county, 1993-1995
Births
(no.)
County
Anencephaly
No. of
cases
Spina bifida
Rate
No. of
cases
Rate
Cameron
Hidalgo
Webb
El Paso
Other 10 counties
23,049
38,264
14,641
46,964
12,303
14
25
14
22
12
6.1
6.5
9.6
4.7
9.8
25
31
11
20
9
10.8
8.1
7.5
4.3
7.3
Total border cases
135,221
87
6.4
96
7.1
All NTDs
Encephalocele
(no. of
cases)
No. of
cases
Rate
95% Clt
4
1
40
62
27
46
22
17.4
16.2
18.4
9.8
17.9
12.4-23.6
12.4-20.8
12.2-26.8
7.2-13.1
11.2-27.1
14
197
14.6
12.6-16.8
1
6
2
* Syndromal NTDs, associated mechanical anomalies, and known chromosomal anomalies are not included in this table,
t Rates are cases per 10,000 live births.
t Cl, confidence interval.
Am J Epidemiol
Vol. 149, No. 12, 1999
1122
Hendricks et al.
bifida rates of 6.4 and 7.1 per 10,000 and an anencephaly :spina bifida ratio of 0.90:1. Eighty-five percent (168 cases) of 197 cases reached a gestational age
of >20 weeks. Including only cases of >20 weeks' gestation in the rate had a more marked effect on reducing
the anencephaly rate (4.9 per 10,000) than the spina
bifida rate (6.7 per 10,000) and decreased the anencephaly:spina bifida ratio to 0.7:1.
Pregnancy outcomes by NTD type and gestational
age are shown in table 2. Eighteen of 87 (21 percent)
anencephaly cases and four of 96 (4 percent) spina
bifida cases were electively aborted at a gestational
age of less than 20 weeks. Another 16 (18 percent)
anencephaly cases and 10 (10 percent) spina bifida
cases were electively aborted from the gestational age
of 20 through 33 weeks. Seventy-three of 76 (96 percent) spina bifida cases and 21 of 28 (75 percent) anencephaly cases that reached the gestational age of 34
weeks were born live. Overall, an NTD was diagnosed
for 75 percent of the cases at some point prior to delivery or abortion. Spontaneous abortions accounted for 3
percent (five cases) and stillbirths for 13 percent (26
cases) of the total NTD cases.
Case-women whose last normal menstrual periods
were from October 1, 1992, to September 30, 1994,
were evaluated for possible seasonal variation of anencephaly, spina bifida, and total NTD prevalence. For
all NTDs the 3-month period that accounted for the
largest number of last normal menstrual periods was
July to September (p < 0.1) (ratchet circular scan test
for a short seasonal peak), with a trough (p = 0.052) 6
months after an August 24 peak date (Edward's test)
(31). Data for live births were also seasonal: 27 percent
of the 90,266 live births for the relevant period 9
months later, July 1, 1993, through June 30, 1995,
occurred during August through October (p < 0.005)
with a trough 6 months after September 18. After
adjustment for the strongly seasonal live birth denominator data, the trough for the last menstrual period of
case-women (all NTDs) occurred 6 months after the
peak date, August 12 (p = 0.31). When anencephaly
and spina bifida data were analyzed separately, only
the last normal menstrual periods for women with
anencephaly-affected pregnancies displayed a significant peak date and a deviation from uniform incidence
(p < 0.05). The significance of the peak date (August
3; p = 0.02) was lost after adjustment for the denominator data (July 23; p = 0.08). No seasonality was
observed for the last normal menstrual periods of
women with spina bifida-affected pregnancies.
The mean maternal age was 24.7 (standard deviation, 6.6) years. Fifty-seven percent of the cases
occurred among women aged <24 years. A U-shaped
distribution of rates by maternal age was observed
with maxima for women aged <19 years (21.4 per
10,000) and >35 years (18.9 per 10,000). Thirty-four
percent of the cases were primipara, 46 percent were
para II or III, and 20 percent were para >IV. Women
with parity >IV had a higher rate (16.2 per 10,000)
than primiparas with a rate of 14.0, but this difference
was not significant (p = 0.46). Phenotypic sex was
ascertained for 179 (91 percent) fetuses or infants.
Overall, female NTDs were more common (62 percent) than male (38 percent). The male:female sex
ratios were 0.54 for anencephaly, 0.68 for spina bifida,
and 0.86 for encephalocele.
Myelomeningoceles (spina bifida with protrusion of
membranes and spinal neural elements into a cyst) comprised 73 (76 percent) of the 96 spina bifida cases and
37 percent of the 197 NTDs. They were distributed as
follows: cranial (n = 1; 1 percent), cervical (n = 4; 6 percent), thoracic (n = 18; 25 percent), lumbar (n = 43; 59
percent), sacral (n = 1; 1 percent), and unspecified (n =
6; 8 percent). Fifty-seven (78 percent) of the myelomeningoceles were associated with hydrocephalus; 86
percent of these were in the thoracolumbar area.
Table 3 details the differences in case definition and
exclusion criteria for selected NTD surveillance sys-
TABLE 2. Pregnancy outcome by neural tube defect type and gestational age, Texas, 1993-1995
Gestational age
Pregnancy
outcome
<20t
20-33
£34
All ages
0(0)*
18(86)
0(0)
7(23)
16(52)
8(25)
21 (60)
0(0)
14 (40)
28 (32)
34(39)
22 (26)
3(14)
0(0)
0(0)
35(100)
Live birth
Elective abortion
Stillbirth
Spontaneous
abortion
Total
21 (100)
31 (100)
Total*
Spina bifida
Anencephaly
£34
All ages
0(0)
4(67)
0(0)
3(20)
10(67)
2(13)
73 (98)
1(1)
1(1)
76 (79)
15(16)
3(3)
112(57)
54(27)
26(13)
3(3)
2(33)
0(0)
0(0)
2(2)
5(3)
87(100)
6(100)
75(100)
96(100)
197(100)
<20
20-33
15(100)
* Total includes 14 encephaloceles: eight were born live, five were electively aborted, and one was stillborn,
t Weeks of gestation are weeks from last normal menstrual period.
t Numbers in parentheses, percentage.
Am J Epidemiol
Vol. 149, No. 12, 1999
Neural Tube Defects along the Texas-Mexico Border
1123
TABLE 3. Case definitions and exclusion criteria for selected neural tube defect (NTD) surveillance
systems in the United States
Case definition
Texas-Mexico border counties
South Carolina
Metropolitan Atlanta
California
Exclusion criteria
Period
Gestational
age
A+
SB + E*
1993-1995
1993-1995
1990-1991
1990-1991
1989-1991
All
All
>20 weeks
Adjusted
All
X
X
X
X
A + SBf
Recognized Syndromal
causes^
NTDs
X
X
X
X
X
X
* Anencephalus/craniorachischisis/iniencephaly (A), spina bifida including meningocele (SB), and encephalocele (E).
t Anencephalus/craniorachischisis/iniencephaly (A) and spina bifida without meningocele (SB).
$ Known syndromes, chromosomal anomalies.
tems in the United States. Table 4 compares the TexasMexico border NTD rates with those of the other surveillance systems. If fetuses of less than 20 weeks are
excluded from the NTD rates, the overall TexasMexico border rate is significantly higher (p < 0.003)
than the rates of South Carolina, the MACDP counties
(22), and California (18, 20, 21). If all gestational ages
are included, the overall NTD rate for the TexasMexico border is similar to that of South Carolina (p =
0.99). In contrast, the border NTD rate for all gestational ages was mildly higher than the adjusted rate for
the MACDP counties (p = 0.052) and significantly
higher than an estimated rate for California (p < 0.01)
(21).
Table 5 compares Hispanic and Anglo rates by
region. Neither the Hispanic nor the Anglo rates for
Texas differed significantly from the white rate for
South Carolina or the adjusted white rate for the
MACDP counties. For >20 weeks' gestation, the overall Hispanic rate of 14.9 for Texas was higher than that
of 9.4 for California (no significance test was possible
because published denominator data were lacking).
TABLE 4. Comparison of neural tube defect (NTD) rates, gestational ages, and ratios for Texas border counties and other
geographic areas
Gestational
oouniy
California:):
11 California counties§
Velie and Shaw:):
Metropolitan AtlantaH
Roberts et al.#
South Carolina**,tt
Cameron County, Texas}::):
Texas border
PprinH
r CIIUU
1983-1987
1990-1992
1989-1991
1989-1991
1993-1994
1990-1991
1990-1991
1993-1995
1993-1995
1980-1989
1990-1992
1993-1995
1993-1995
Anencephaly
age
(weeks)
No. of
cases
>20
>20
>20
All
>20
>20
All
>20
All
>20
>20
>20
All
193
58
135
279
13
19
37
33
66
30
24
66
87
Spina bifida
Ratet
No. of
cases
Rate
2.7
2.0
1.9
3.9
1.6
2.5
4.8
2.6
5.2
6.5
12.4
4.9
6.4
272
116
247
352
31
30
39
61
94
38
15
90
96
3.8
4.0
3.5
5.0
3.9
3.9
5.1
4.8
7.3
8.3
7.8
6.7
7.1
Encephalocele
(no. of
cases)
15 (other)
26 (other)
7
10
11
16
27
12
14
NTDs
A:SB*
Total
Rate
465
174
397
657
51
59
87
110
187
72
41
168
197
6.5
6.0
5.6
9.3
6.4
7.7
11.3
8.6
14.6
15.7
21.2
12.4
14.6
0.7:
0.5:
0.5:
0.8:
0.4:
0.6:
0.9:
0.5:
0.7:
0.8:1
1.6:1
0.7:1
0.9:1
* A:SB, ratio of anencephaly to spina bifida.
t Rates are cases per 10,000 live births.
t California Birth Defects Monitoring Program (E. M. Velie and G. M. Shaw. Am J Epidemiol 1996;144:473-9).
§ Birth defects in eleven California counties: 1990-1992. Emeryville, CA: California Birth Defects Monitoring Program, California
Department of Health Services, June 1996.
H Metropolitan Atlanta Congenital Defects Program (L. Paulozzi, National Center for Environmental Health, Birth Defects, and
Developmental Disabilities, Centers for Disease Control and Prevention, personal communication, 1995).
# H. E. Roberts et al. Pediatrics 1995;96:880-3.
** South Carolina Neural Tube Defect Surveillance, Prevention, and Research Initiative (R. Stevenson and J. H. Dean, Greenwood
Genetics Center, personal communication, 1995).
t t 1993 surveillance data are based on 14 counties for 9 months and the entire state for 3 months; 1994 data are for the entire state.
i J. Henry, Texas Department of Health, personal communication, 1995.
Am J Epidemiol Vol. 149, No. 12, 1999
1124
Hendricks et al.
TABLE 5. Comparison of Hispanic and white/Anglo neural
tube defect rates by region
US
region
Period
Texas border
1993-1995
South Carolina
Metropolitan Atlanta
California
1993-1995
1990-1991
1983-1987
Rates
Gestational
age
White/
Hispanic
(weeks)
Anglo
>20
All
All
All*
>20
12.9
14.9
9.0
7.4
10.6
16.3
11.8
5.9
* Adjusted.
DISCUSSION
Using a multisource active surveillance system, this
study was the first to establish baseline NTD rates for
the entire, predominantly Hispanic, Texas-Mexico border and to ascertain NTD-affected pregnancies at all
gestational ages. Rates for each of these years were
remarkably stable for this population, with no significant differences among the 1993, 1994, and 1995 NTD
rates. Moreover, the 3-year 14-county rate of 14.6 per
10,000 live births was not significantly different from
the Cameron County rate of 14.7 per 10,000 for
1986-1989, the period preceding the 1990-1991 cluster (Texas Department of Health, unpublished report,
1992). This suggests that this is a valid baseline rate
and that the exposures that occurred in Cameron
County in 1990-1991 were of a transient nature. The
only significant variation in the border NTD rates for
the 3-year period was a lower NTD rate in El Paso
County compared with that for the rest of the border.
The border rate was somewhat higher than a prenatally adjusted rate for the MACDP counties (p = 0.052)
(22) and the recently estimated rate for California (p <
0.001) but the same as that of South Carolina. A number of factors may account for the intra- and interstate
variability in rates. Hispanic ethnicity and low socioeconomic status are well-accepted NTD risk factors
(10). The Texas-Mexico border population is predominantly Hispanic (82 percent of the population and 91
percent of the live births) (31) with high levels of
poverty (36 percent) (32) and a significant proportion
(22 percent) engaged in migrant or seasonal farm
working (32).
In contrast, the South Carolina and MACDP county
populations are primarily white and black. Even in
California, a much smaller proportion (26 percent) of
the statewide population is Hispanic (33). Poverty
rates for California, South Carolina, and the MACDP
counties are below 20 percent (34-36), and
migrant/seasonal farm workers comprise less than 5
percent of these populations (37). Thus, populations in
California, South Carolina, and the MACDP are less
Hispanic and less poor, and they have much lower percentages of persons engaged in migrant and seasonal
farm work than the border population in Texas. These
same statements are true for the population in El Paso
County (73 percent Hispanic, 27 percent poverty, 1
percent migrant workers) compared with that of the
rest of the Texas-Mexico border. In summary, the geographic variability evident in these data may be attributable to host susceptibility (i.e., genetic polymorphisms), to a variety of nongenetic factors such as
environmental, occupational, or nutritional exposures,
or, most likely, to an interaction between genetic and
nongenetic factors.
The lower NTD rate for California may also be
attributable to differences in case-finding methodologies, completeness of ascertainment, exclusion criteria, or regional variation in prenatal NTD diagnosis
followed by abortion. Although multiple case-finding
sources were used by the Texas-Mexico border NTD
project, only about 5 percent of the Texas cases were
identified solely through nonhospital sources (38).
Because California used some out-of-hospital surveillance sources (18), albeit different from those used in
Texas, this particular methodological difference is
unlikely to have accounted for more than a small portion of the difference between the rates for the TexasMexico border and California.
Surveillance systems that include only infants and
fetuses of >20 weeks' gestational age underrepresent
the true number of birth defects (23). This is particularly likely for birth defects, such as anencephaly, in
which prenatal diagnosis is often followed by an elective abortion at <20 weeks. When restricted to gestational ages of >20 weeks, the Texas-Mexico border
rate was much higher than rates in South Carolina,
MACDP, and California. For Texas Hispanics, a traditionally Catholic ethnic group, it is likely that the religious proscription against abortion was a major factor
in the high percentage of NTD-affected pregnancies
carried to term. In a recent study that compared surveillance systems in six states over different time periods, the percentage of electively aborted anencephaly
and spina bifida cases ranged from 9 to 42 percent
(23). In one Midwestern tertiary care center, 100 percent of prenatally diagnosed anencephalics and 78 percent of prenatally diagnosed spina bifida-affected
pregnancies were aborted (39). In the same
Midwestern study, 82 percent of the women who electively chose to abort did so before 20 weeks (39). For
the Texas-Mexico border counties, 27 percent of the
anencephaly and spina bifida cases were electively
aborted; 45 percent aborted at less than 20 weeks.
Variability in NTD rates among studies may also be
due to differences in case definitions or eligibility criAm J Epidemiol Vol. 149, No. 12, 1999
Neural Tube Defects along the Texas-Mexico Border
teria. Studies often include only nonsyndromal NTDs
(40). Some exclude encephaloceles entirely (18, 21),
and others have grouped encephaloceles with either
spina bifida or anencephaly (10). With the exception of
encephaloceles, most of these disorders account for so
few cases that their inclusion or exclusion has little
effect on a rate. In contrast, exclusion of encephaloceles that account for a much larger proportion of
NTDs may decrease a rate by 10 percent or more.
The similarity of the Texas-Mexico border rates to
the South Carolina rates is rather striking. The thoroughness of case ascertainment and the inclusion of all
gestational ages probably account for some of the
comparability. However, the similarity of these rates
despite greatly different ethnic populations remains
difficult to explain. Perhaps a folic acid deficiency is
common to all. Other nutritional exposures also may
be of some interest. Corn, in the form of tortillas and
grits, figures prominently in both the Texas-Mexico
border and South Carolina diets. A class of mycotoxins
in corn, fumonisins, was linked to fatal outbreaks in
1989 and 1990 of equine leukoencephalomalacia (41,
42) and porcine prenatal and neonatal mortality and
pulmonary edema (42,43). Although the effects on the
human fetus are unknown, recent in vitro experiments
have shown that fumonisin Bj inhibits folate receptormediated transport of 5-methyltetrahydrofolate in both
a time- and concentration-dependent fashion (44). To
our knowledge, the only transient "environmental"
exposure coincident with the Cameron County cluster
was the fumonisin-contaminated corn crop harvest
responsible for the above-mentioned epizootics (42).
The 0.9:1 anencephaly:spina bifida ratio for the
Texas border is higher than ratios (except for NTD
clusters) for other parts of the United States (23) but
much lower than the 2.5:1 ratio found during the
1990-1991 Cameron County cluster (Texas
Department of Health, unpublished report, 1992). The
anencephaly:spina bifida ratios reported for small
NTD clusters have generally reflected increases in
either the anencephaly or spina bifida rate but not both
(44—48). Some authors have posited that environmental and occupational exposures play more dominant
roles in the etiology of anencephaly than in that of
spina bifida; this would be reflected in the anencephaly :spina bifida ratio (1, 49). This, again, suggests
that there may have been some sort of "environmental"
exposure in Cameron County during 1990 and 1991
that failed to persist in subsequent years.
Many findings in this study are consonant with those
of earlier studies. Maternal age at delivery or abortion
displayed the expected U-shaped distribution of rates
(15). Although the correlation between parity and NTD
risk has been variously described as both U-shaped and
Am J Epidemiol
Vol. 149, No. 12, 1999
1125
linear (15), this study demonstrated a mild but nonsignificant risk with increasing parity. In contrast to a
recent California study (18), the expected predominance
of phenotypically female NTDs (3, 49) was observed in
this study. (Some phenotypic females are genotypic
males. R. Stevenson, Greenwood Genetics Center, personal communication, 1996). The male:female NTD
ratios were 0.63 for all NTDs, 0.54 for anencephaly, and
0.68 for spina bifida. As has been noted previously,
there was a female excess (75 percent) of "high" spina
bifida defects (thoracic level or higher), and there was a
1:1 ratio of male:female "low" defects (lumbar or
lower) (18,50). Both the distribution of myelomeningoceles by level and the frequency of associated hydrocephalus have been observed previously (50). After
adjustment for a live birth peak that occurred in early
fall, only a mild late summer peak remained for anencephaly conceptions (p = 0.08). Neither spina bifida nor
total NTD conceptions showed significant seasonality.
As was shown previously in California (18), Mexicoborn Hispanic case-women had a significantly (p <
0.01) higher NTD rate (15.1 per 10,000) than United
States-born Hispanic case-women (9.5 per 10,000).
Some underascertainment of cases undoubtedly
occurred in this study; cases may have been missed for
pregnancy outcomes occurring in counties or states
outside the study area. This is most likely for the two
southernmost counties, in which more than 39 percent
(32) of the population is engaged in migrant and seasonal farm work. It is also possible that medical
records may have been miscoded so that cases were
not identified by the ICD-9 screening codes. However,
multisource case finding makes this unlikely. On average, most cases identified within hospitals were found
in 1.1 additional non-ICD-9 sources (especially prenatal clinics, genetic centers, and radiology centers);
only 12 percent were found exclusively through ICD-9
codes (38). The specific type of NTD may have been
misclassified in a medical record. Finally, the outcome
of the NTD-affected pregnancy may occasionally have
been assigned to the wrong county. Determination of a
mother's residence at the time of birth or abortion was
often difficult because of migration for birth to and
from the border counties and to and from Mexico.
The rates reported in this paper are surveillance
rates. The numerators reflect the residence reported by
the case-mother on the medical record for all types of
delivery outcome. Likewise, the denominators reflect
the residence recorded on the birth certificate as
reported to vital records. Some misclassification bias
of both the numerator and denominator is likely.
Because the Texas Neural Tube Defect Project also
includes a folic acid intervention, it was known
through follow-up that 13 (6.6 percent) of the 197
1126
Hendricks et al.
case-women, whose medical records recorded their
residency as one of the 14 border counties, actually
resided in Mexico. Although it is impossible to adjust
the denominator, censoring these 13 case-women from
the numerator would reduce the overall NTD rate by
6.8 percent, from 14.6 to 13.6 per 10,000 live births (p =
0.51).
The strengths of this study, which established a
baseline NTD rate for the Texas-Mexico border, are
the use of multisource case finding (nine out-ofhospital sources were used) and the lack of exclusion
criteria for gestational age. The four state populations that are compared appear to have similar case
definitions and exclusion criteria, and all now
include NTD-affected outcomes at <20 weeks of gestation. The significantly higher rates noted for Texas
and South Carolina compared with that for California
can be interpreted in only one of two ways. Either
Texas and South Carolina have unique high-risk populations or their case-finding sources and methodologies are more inclusive. In addition to surveillance,
the Texas-Mexico border project offers women identified through the project education, folic acid supplementation, and the option of participating in a
case-control study. A case-control study is under way
to determine whether the high rates for the border
population are due to genetic or nongenetic (e.g.,
environmental or nutritional) risks or are merely a
reflection of improved case finding and lack of
exclusion criteria.
ACKNOWLEDGMENTS
Supported in part by the US Department of Health and
Human Services, Public Health Service, Centers for Disease
Control
and Prevention,
cooperative
agreement
U85/CCU608761.
The authors thank the following NTD team members for
their crucial role in identifying the case-women and for
enrolling them in the prevention program: El Paso: Hilda
Chavarria, Maria Torres, Carmen Ramos, Donna Brom, and
Patricia Velazquez; Harlingen: Oralia Villafranca, San
Juana Thompson, Graciela Rubio, Manuela Flores, Rene
Rodriguez, Sara Mungia, and Jorge Trevino; and Laredo:
Ricardo Trevino, Miguel Madrigal, Olivia Macias
Gutierrez, Cynthia Medina de Llano, Jackie Bassini, and
Armandina Ortiz.
The authors are also grateful to Drs. Lucina Suarez, Jean
Brender, and Mark Canfield for their insightful comments
and to Xuan-Nguyen Huynh, Mardi VanEgdom, Leland
Carmichael, and Sue Gadd for their invaluable assistance.
Finally, they wish to acknowledge the dedication and support of the labor and delivery and medical records staffs of
the 21 delivery hospitals in the study area.
REFERENCES
1. Van Allen MI, Kalousek DK, Chernoff GF, et al. Evidence for
multisite closure of the neural tube in humans. Am J Med
Genet 1993;47:723-43.
2. Golden JA, Chernoff GF. Multiple sites of anterior neural tube
closure in humans: evidence from anterior neural tube defects
(anencephaly). Pediatrics 1995;95:506-10.
3. Copp AJ, Brook FA, Estibeiro JP, et al. The embryonic development of mammalian neural tube defects. Prog Neurobiol
1990;35:363-403.
4. Recommendations for the use of folic acid to reduce the number of cases of spina bifida and other neural tube defects.
MMWR Morb Mortal Wkly Rep 1992;41(RR-14):l-7.
5. Nau H. Valproic acid-induced neural tube defects. In: Bock G,
Marsh J, eds. Neural tube defects. Ciba Foundation symp 181.
Chichester, England: John Wiley & Sons, Ltd, 1994:144-60.
6. Warkany J. Aminopterin and methotrexate: folic acid deficiency. Teratology 1978;17:353-8.
7. Shaw EB. Fetal damage due to maternal aminopterin ingestion. Am J Dis Child 1972; 124:93^.
8. Mills JL, Baker L, Goldman S. Malformations in infants of
diabetic mothers occur before the seventh gestational week.
Diabetes 1979;28:292-3.
9. Holmes LB. Spina bifida: anticonvulsants and other maternal
influences. In: Bock G, Marsh J, eds. Neural tube defects. Ciba
Foundation symp 181. Chichester, England: John Wiley &
Sons, Ltd, 1994:232-8.
10. Elwood JM, Little J, Elwood JH. Epidemiology and control of
neural tube defects. Oxford, England: Oxford University
Press, 1992.
11. Milunsky A, Ulcikas M, Rothman K, et al. Maternal heat exposure and neural tube defects. JAMA 1992,268:882-5.
12. van der Put NMJ, Steegers-Theunissen RPM, Frosst P, et al.
Mutated methylenetetrahydrofolate reductase as a risk factor
for spina bifida. Lancet 1995;346:1070-1.
13. Ou CY, Stevenson RF, Brown VK, et al. C677T homozygosity associated with thermolabile 5,10-methylenetetrahydrofolate reductase as a risk factor for neural tube defects.
(Abstract). In: Byers P, Wilkes R, Basin P, et al, eds. Program
and abstracts of the 1995 annual meeting of the American
Society of Human Genetics. Chicago, IL: University of
Chicago Press, 1995.
14. Greer K, Barber R, Shaw G, et al. Folate receptor gene variants. (Abstract). Presented at the 3rd Joint Clinical Genetics
Meeting of the 27th Annual March of Dimes Clinical Genetics
Conference and the American College of Medical Genetics 3rd
Annual Meeting, San Antonio, Texas, March 11-14, 1996:94.
15. Little J, Elwood JM. Epidemiology of neural tube defects. IV.
Migrants and their descendants. In: Kiely M, ed. Reproductive
and perinatal epidemiology. Boca Raton, FL: CRC Press, Inc,
1991.
16. World Health Organization. The international clearing house
for birth defects monitoring systems. Congenital malformations worldwide. Amsterdam: Elsevier Science Publishers, B V,
1991.
17. Strassburg MA, Greenland S, Portigal LD, et al. A populationbased case-control study of anencephalus and spina bifida in a
low-risk area. Dev Med Child Neural 1983;25:632-41.
18. Shaw GM, Jensvold NG, Wasserman CR, et al. Epidemiologic
characteristics of phenotypically distinct neural tube defects
among 0.7 million California births, 1983-1987. Teratology
1994;49:143-9.
19. Yen IH, Khoury MJ, Erickson JD, et al. The changing epidemiology of neural tube defects. United States, 1968-1989.
Am J Dis Child 1992;146:857-61.
20. Birth defects in eleven California counties: 1990-1992.
Emeryville, CA: California Birth Defects Monitoring Program,
California Department of Health Services, June 1996.
21. Velie EM, Shaw GM. Impact of prenatal diagnosis and elective
termination on prevalence and risk estimates of neural tube
defects in California, 1989-1991. Am J Epidemiol 1996;144:
Am J Epidemiol Vol. 149, No. 12, 1999
Neural Tube Defects along the Texas-Mexico Border
473-9.
22. Roberts HE, Moore CA, Cragen JD, et al. Impact of prenatal
diagnosis on the birth prevalence of neural tube defects,
Atlanta, 1990-1991. Pediatrics 1995;96:880-3.
23. Surveillance for anencephaly and spina bifida and the impact
of prenatal diagnosis—United States, 1985-1994. MMWR
CDC Surveill Summ 1995,44:1-13.
24. Brender JD, Carmichael L, Preece MJ, et al. Epidemiology of
anencephaly in Texas, 1981-1986. Tex Med 1989,85:33-5.
25. Brender JD, Suarez L. Paternal occupation and anencephaly.
Am J Epidemiol 1990;131:517-21.
26. Canfield MA, Annegers JF, Brender JD, et al. Hispanic origin
and neural tube defects in Houston/Harris County, Texas. Am
J Epidemiol 1996;143:1-11.
27. Lynberg M, Edmonds LD. Surveillance of birth defects. In:
Halperin W, Baker E, eds. Public health surveillance. New
York: Van Nostrand Reinhold, 1992.
28. Khoury MJ, Edmonds LD. Metropolitan Atlanta Congenital
Defects Program: twenty-five years of birth defects surveillance at the Centers for Disease Control. In: Mastroiacovo P,
ed. Proceedings of the Third ASM International Symposium
on Birth Defects. Rome, Italy: ICARO-ASM Press, 1992.
29. Dean AG, Dean JA, Coulombier D, et al. A word processing,
database, and statistics system for epidemiology on microcomputers (Epi Info), version 6. Atlanta, GA: Centers for
Disease Control and Prevention, 1994.
30. Gahlinger PM, Abramson JH. Computer programs for epidemiologic analysis. Honolulu, HI: Makapuu Medical Press,
1993.
31. Goldman DA. The EPIGRAM computer program for analyzing mortality and population data sets. Public Health Rep
1994;109:118-24.
32. The University of Texas-Pan American, Texas-Mexico Border
Health Coordination Office. Texas-Mexico border county
demographics and health statistics: 1994. Edinburg, TX: The
University of Texas System, 1994.
33. US Bureau of the Census. Statistical abstract of the United
States: 1995. 115th ed. Austin, TX: The Reference Press,
1995:36.
34. Akioka LM, ed. Georgia statistical abstracts 1994-1995. Table
1.122. Population of Georgia by race, sex, and county, 1992.
Athens, GA: Selig Center for Economic Growth, University of
Georgia, 1992.
35. US Bureau of the Census. Statistical abstract of the United
States: 1995. 115th ed. Austin, TX: The Reference Press,
1995:482.
36. Bachtel DC, Boatright SR, eds. The Georgia county guide,
1995. 14th ed. Economics. Poverty characteristics, 1989.
Am J Epidemiol Vol. 149, No. 12, 1999
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
1127
Athens, GA: Cooperative Extension Service, University of
Georgia, 1995.
Migrant Health Program. An atlas of state profiles which estimate the number of migrant and seasonal farmworkers and
members of their families. Table II. Adjusted state profiles
(farmworkers plus dependents). Rockville, MD: Department
of Health and Human Services, 1990.
Pae AS. A descriptive analysis of the validity, effectiveness,
and efficiency of the Texas Department of Health ICD-9-CM
screening codes for neural tube defect case ascertainment.
(MPH thesis). Houston: School of Public Health, The
University of Texas Health Science Center at Houston, 1995.
Grevengood C, Shulman LP, Dungan JS, et al. Severity of
abnormality influences decision to terminate pregnancies
affected with fetal neural tube defects. Fetal Diagn Ther 1994;
9:273-7.
Myrianthopoulos NC, Melnick M. Studies in neural tube
defects. I. Epidemiologic and etiologic aspects. Am J Med
Genet 1987;26:783-96.
Ross PF, Rice LG, Reagor JC, et al. Fumonisin Bt concentrations in feeds from 45 confirmed equine leukoencephalomalacia cases. J Vet Diagn Invest 1991;3:238^».
Ross PF, Nelson PE, Richard JL, et al. Production of fumonisins by Fusarium moniliforme and Fusarium proliferatum isolates associated with equine leukoencephalomalacia and a pulmonary edema syndrome in swine. Appl Environ Microbiol
1990;56:3225-6.
Bane DP, Neumann EJ, Hall WF, et al. Relationship between
fumonisin contamination of feed and mystery swine disease. A
case-control study. Mycopathologia 1992;117:121—4.
Stevens VL, Tang J. Fumonisin B^induced sphingolipid depletion inhibits vitamin uptake via the glycosylphosphatidylinositol-anchored folate receptor. J Biol Chem 1997;272:18020-5.
Sever LE, Sanders M, Monsen R. An epidemiologic study of
neural tube defects in Los Angeles County I. Prevalence at
birth based on multiple sources of case ascertainment.
Teratology 1982;25:315-21.
Sever LE. An epidemiologic study of neural tube defects in
Los Angeles County II. Etiological factors in an area with low
prevalence at birth. Teratology 1982;25:323-34.
Sever LE. The conundrum of birth defects clusters. Health
Environ Digest 1994;7:4-6.
Sever LE. Looking for causes of neural tube defects: where
does the environment fit in? Environ Health Perspect 1995;
103(suppl 6): 165-71.
Seller MJ. Neural tube defects and sex ratios. Am J Med Genet
1987;26:699-707.
Akar Z. Myelomeningocele. Surg Neurol 1995;43:113-18.