Spatial distribution of orofacial cleft defect births in Harris County, Texas, 1990 to 1994, and historical evidence for the presence of low-level radioactivity in tap water.

Background: While both ionizing and nonionizing radiation are known to impair human reproductive capacity, the role of low-level domestic radiation continues to be an unsettled issue.

Objectives: We examined the geostatistical distribution (residential longitude and latitude) of orofacial cleft birth cases adjusted for the underlying population distribution. Furthermore, we examined the cleft birth rates enumerated by zip codes for possible associations with levels of radium and radon in drinking water.

Methods: Cleft births and unaffected live births in Harris County, Texas, from 1990 to 1994, were geocoded by residential addresses and tested for spatial clusters using the space-time clustering program SaTScan. Historical sample data on local variations in water quality facilitated the assessment of the association of orofacial cleft defect births with low-level radiation exposure.

Results: A cluster of significantly greater than expected numbers of cleft defect births was identified in northwest Harris County, (relative risk = 3.0, P = 0.043), where the presence of elevated levels of radium (>3 pCi/L) and radon (>300 pCi/L) in the tap water has been known since the 1980s.

Conclusions: Despite the ecological design of the study, lacking individual exposure measurements for cleft birth residences, there was strong suggestive evidence of an association between elevated radiation levels in tap water and elevated cleft birth prevalence rates by zip codes. Attention of physicians is invited to environmental causes as potential risk factors for orofacial cleft. This would aid in genetic counseling and the development of future preventive measures.

Key Words: orofacial cleft, low-level radiation, drinking water, geographic information system (GIS) methods.

**********

The term "facial clefting" refers to a spectrum of defects which involve the upper lip, the palate, or both. (1) The exact cause(s) of orofacial cleft development has not been resolved. (2,3) It is currently theorized that facial clefting is a heterogeneous process resulting from a variety of causes, including genetic factors (4-8) and environmental insults, (9-13) with possible environmental and genetic interactions. (14,15) For the most part, the studies of interactions between genetic and environmental factors have been confined to parental tobacco smoking, (8,9,16-21) and had varying success. Other recent studies highlighted several possible susceptibility loci for low serum folate (22) and low vitamin [B.sub.6] and [B.sub.12] status. (22,23) Orofacial cleft N-acetyltransferase phenotype, maternal smoking, and medication use was also investigated, with negative results. (24) In all, it still remains to be seen whether these factors and their interactions are a sufficient explanation for cleft etiology, (25,26) and the need for further research is apparent. Understanding environmental triggers is important, because the exposures are preventable. (27)

Many authors agree that in utero exposures to ionizing radiation involve teratogenic risks. (27-46) Studies of atomic bomb survivors in Hiroshima and Nagasaki, Japan, indicated that the risk of malformations was greatest between 8 to 15 weeks of gestation. (33,38) The major defect in the children born after the A-bomb was a decrease in head size accompanied with mental retardation. (47) Studies in the former East Germany reported a 9.4% increase in the rates of cleft deformities between 1987 and 1989, subsequent to the meltdown of the Chernobyl reactor in 1986. (32,35)

Brent (29) reported growth retardation, microcephaly, and mental retardation in the offspring of women, who were treated with therapeutic radiation for various medical reasons and before knowing that the patient was pregnant. Ritenouer (31) found that maternal exposure to doses above 15 rems early in pregnancy resulted in preimplantation death, growth retardation, and central nervous system defects.

Domestic-level radiation has been implicated in the development of birth defects, including cleft defects, as far back as the 1950s. (42) A study in New York State found an elevated rate of malformations in areas underlain by uranium-bearing rocks. Rates of orofacial cleft were 33% greater than in areas classified as unlikely for exposure to uranium decay products. However, no data on actual concentrations of radioactive isotopes in the drinking water supplies were available. This was a weakness of an otherwise pioneering study.

[FIGURE 1 OMITTED]

In all, while both ionizing and nonionizing radiation were shown to impact human reproductive capacity and produce birth defects, with few exceptions, the data were derived from studies involving cases of acute exposure and high doses. There are limited data available regarding reproductive and teratological consequences associated with domestic levels of radiation exposures and an obvious need to reduce the uncertainty. (48)

Present Study

During the 1980s and early 1990s, elevated concentrations of radium-226 and radon-222 in groundwater wells and tap water were reported in Harris County, Texas, particularly in the northwest. (49-55) Radium concentrations, as great as 15 pCi/L, were found in the water, in an area near a major hydrocarbon field. Radon concentrations, as great as 2,000 pCi/L, were also reported in the water (Figs. 1 and 2). In contrast, the part of Harris County supplied by surface water had no radionuclide concentrations above acceptable standards. (56) The availability of these historical data and the presence of regional variations in concentrations of radium and radon in tap water created an opportunity in the present study to assess the association of cleft birth defects with low-level radiation exposure.

[FIGURE 2 OMITTED]

Methods

This was a medical-geographic study of the distribution of cleft palate and cleft lip incidence in infants born to families residing in Harris County, Texas, between 1990 and 1994. Texas birth certificates were the source of cleft palate/cleft lip deformity information. Residential addresses of cases and live births were geocoded by longitude and latitude, using StreetInfo and MapMarker software by MapInfo, Inc., (57,58) (Fig. 3).

To investigate the spatial clusters of orofacial births in Harris County, Texas, we used the space-time clustering program SaTScan. (59) A longitude and latitude grid for Harris County was constructed containing square "blocks" measuring 0.01 degrees (approximately 0.6 miles) on each side. This allowed counting the number of cases versus noncases within each "block." The "block" locations were modeled using a Bernoulli clustering algorithm, which progressively grouped excess density clusters and adjusted the spatial density of the cases for the underlying inhomogeneity of the background population (Fig. 4).

The orofacial defect births were aggregated by the Harris County zip codes to calculate rates per 10,000 live births. The independent variables were the previously reported radionuclide concentrations in water supplies in Harris County, collected during the period 1987 to 1992. (Figs. 1 and 2). (50-55) Recently obtained data for the period 1999 to 2002 from the Texas Commission for Environmental Quality (TCEQ) were geocoded and plotted by the authors (data not shown), indicating that the radiation levels in the well water had generally not diminished.

Concentrations of radium-226 were dichotomized as "<3 pCi/L" and "[greater than or equal to]3 pCi/L." The current maximum contaminant level for total radium is 5 pCi/L (isotopes of Ra-226 and Ra-228 combined). Concentrations of radon-222 were dichotomized as "<300 pCi/L" and "[greater than or equal to]300 pCi/L," the latter being the limit proposed by the Environmental Protection Agency for public water supplies. (60) Zip codes corresponding to either category of potential exposure were appropriately indexed. The boundaries of Harris County 1990 zip codes were available from MapInfo.

The differences in rates of cleft birth defect by zip codes located in the areas classified as <3 pCi/L and [greater than or equal to]3 pCi/L of radium-226 were compared (Table 1). Analyses were repeated for the radon-222 dichotomies of <300 pCi/L" and [greater than or equal to]300 pCi/L (Table 2). Multivariate analyses of the dichotomies were repeated, adjusting zip code rates for potentially confounding variables, including (a) population density, (b) racial and ethnic composition, (c) per capita income, (d) level of educational achievement, and (e) ratio of white-to-blue collar workers. The data for 1990 zip codes were obtained from TargetPro digital database available from MapInfo.

[FIGURE 3 OMITTED]

Results

There were 167 births with orofacial defects with a reported residential address in Harris County for the period 1990 to 1994, resulting in a mean annual rate of 5.7 per 10,000 live births. Results of a Bernoulli spatial analysis identified a cluster of 22 cases, compared with a background noncase population of 12,588, centered at -29.969 N, 95.560 W, with a radius of approximately 5 miles (8 km). On the Harris county map, this cluster coincided with the intersection of four major highways: 290 and 249 going out of Houston to NW, and highway 6 and beltway 8 encircling Houston from the NW. This area is marked on Figure 4 as an area of significantly increased risk for cleft defect births. The expected number of cases, adjusted for the number of live births in this area, was 7.3, resulting in a relative risk of 3.0 (95% CI (1.8, 4.3), P = 0.043). The clustering algorithm did not identify any other clusters where the relative risk of the adjusted case density significantly exceeded that of noncases.

Zip codes with the highest rates of cleft defect births (up to 39.5/10,000 live births) in northwest Harris County spatially coincided with the area where elevated concentrations of Ra-226 and Rn-222 have been known to occur since the 1980s. Table 1 shows that the mean annual rates for cleft defect births (14.6/10,000) for areas with Ra-226 [greater than or equal to]3 pCi/L were 2.70 times greater (95% CI 1.84, 3.54) compared with 5.4/10,000 for areas with Ra-226 <3 pCi/L. This difference was highly significant (P < 0.001). We subsequently adjusted cleft rates for potential confounding variables. None of the covariables were significantly associated with cleft rates and the adjustment failed to diminish the statistical significance of radium concentrations in domestic water.

Table 2 shows that the mean annual rates for cleft defect births in Harris County were 9.3/10,000 live births in zip codes which were classified as having elevated radon 222. This was 1.7 times greater (95% CI 1.08, 2.31) compared with 5.5/10,000 in zip codes with Rn-222 concentrations below 300 pCi/L. This difference was significant (P < 0.05) and, as with radium, the adjustment for potentially confounding variables did not diminish the significance of radon concentrations.

Discussion

The median rate of cleft birth defects observed for Houston/Harris County, Texas, for the period between 1990 and 1994, was 4 per 10,000 live births and the mean rate was 6 per 10,000 live births. These statistics were in the same range as rates reported by other investigators (2,15,61) However, within the study area, rates of cleft defects exhibited significant variation, with a cluster of elevated rates in the northwest, at the intersection of four major highways.

[FIGURE 4 OMITTED]

Two to three decades ago, the area where the spatial cluster of cleft defect births was observed was sparsely populated, and the principal activities in northwest Harris County traditionally were rice farming and hydrocarbon exploration. In the late 1980s and 1990s, urban development began spreading rapidly into this area and this growth continues. Between 1991 and 2000, Harris County, Texas, experienced a 20.7% increase in population, and the "Houston Energy Corridor" extended north-westward. (62)

Due to this urban expansion, many groundwater wells were drilled in northwest Harris County to accommodate the need for public and private water supplies. Some of the previous land uses might not have been fully compatible with this urban advance. It was discovered, for instance, that the benzene concentration in one of the public water wells in northwest Harris County, south of the City of Tomball, was 11 times greater than the maximum allowed under the Safe Drinking Water Act. (63,64)

Furthermore, elevated concentrations of radium-226 and radon-222 were reported in some municipal utility districts in this part of Harris County as far back as 20 years. Data gathered by Brock (49) in 1984 indicated that at least 12 municipal utility districts in northwest Harris County violated standards with respect to radium in the public drinking water. Cech and coworkers gathered further data on the distribution of radium-226 in the water and added radon-222 measurements in Harris County and other parts of Texas. (50-55)

Gentry and coworkers (42) in New York State reported that cleft birth defect rates varied, depending on the likelihood of radioactivity being present in the drinking water. The rates of cleft defects ranged from 12 to 16 per 10,000 live births, with the lower end of the range observed in areas classified as "unlikely" for radiation exposure via water, and the higher end of the range observed in areas classified as "probable" for radiation exposure via water.

In the present study in Harris County, Texas, the mean annual incidence rate of cleft defects was 15 per 10,000 live births for a group of zip codes situated in the area with Ra-226 concentration in drinking water [greater than or equal to]3 pCi/L. This was significantly greater than 5 per 10,000 live births observed in parts of Harris County with an unlikely waterborne exposure to this isotope. The higher end of this range was consistent with that reported in New York State.

The mean annual rate of cleft birth defects in Harris County was 9 per 10,000 live births for a group of zip codes with elevated concentration of another uranium progeny, Rn-222. This was at least 1.5 times greater than 6 per 10,000 live births or less observed in parts of Harris County with no appreciable concentrations of radon. The differences in cleft rates dichotomized for both radium and radon concentrations were highly significant (P < 0.001 and P < 0.05). These differences persisted after adjustment for other factors, such as differences in per capita income, educational achievement, proportion of white-to-blue collar occupations, and ethnicity.

It is biologically plausible that radioisotopes of uranium (Ra-226 and Rn-222) were a contributing factor in the development of cleft birth anomalies in northwest Harris County. Radiation is known to produce spermatotoxic and mutagenic effects. (36) Dixon (65) included both ionizing and nonionizing radiation among agents capable of impairing human reproductive capacity, both male and female. Exposure to ionizing radiation was reported to produce microcephaly, mental retardation, growth retardation, and Down syndrome. (66) Animal studies (44) also reported developmental malformations, including meningoencephalocele, spina bifida, eye defects, tail defects, and edema in irradiated laboratory mice. Researchers interpreted the mechanism of radiation-induced teratogenesis as resulting from direct cell destruction, with residual DNA damage.

In our opinion, there is another mechanism by which uranium decay products may contribute to structural birth defects. Radium is a known "bone seeker." (67-70) Being biologically analogous to calcium, radium competes for calcium sites in the growing fetal bone matrix. According to Mays and coworkers, (69) most of the radium atoms are deposited in the growing areas, which makes the developing embryo particularly sensitive.

Radon-222 gas is the first progeny of decaying radium. Mays and coworkers (70) measured radon retention in the bones of radium-injected beagles. They reported that practically all radon retention occurred in bone crystals and not in fat or other organic material, and that radon recoiled during radium disintegration, which caused damage to the bone structure. Significantly, Mays and coworkers have shown that the highest fractional radon retention (up to 22%) occurred in the jawbone and calvarium. These authors indicated that fractional radon retention after exposure was similar in rats, dogs, and man.

An important paper by Purnell and coworkers (37) pointed out that in early gestation, the fetal skeletal structure is primarily cartilaginous, and calcium uptake increases markedly during the bone formation process. Radon-222 is too short-lived to be significantly transferred across the placenta to the fetus. However, radon decay progenies, isotopes of lead (Pb-210) and polonium (Po-210), will follow calcium into the fetal skeleton during maturation. This mechanism may explain defective fusion of skeletal components, including possibly cleft and other structural defects.

In our investigation using geospatial methods, we recognized the importance of scale. Geographical referents were assigned by latitude and longitude to addresses of cleft defect births, which allowed testing for spatial clusters using SaTScan. In subsequent steps, rates per 10,000 live births were calculated for 1990 zip codes and compared by zip codes with the likely versus unlikely presence of radiation in tap water. Further analyses were adjusted for potentially confounding effects of demographic variables, also by 1990 zip codes. In a parallel study carried out by our group, cleft defect births were aggregated by 1990 census tracks instead of zip codes. This produced similar results, albeit with greater variation in the incidence rates due to smaller numbers of both cleft defect births and overall live births in census tracks compared with zip codes. This study will be detailed in a subsequent communication.

Today, Texas has one of the best birth defect registries in the country, but the registry has a relatively recent history. At this writing, birth defect registry data in Harris County were available for only 3 years, from 1999 to 2002. This registry was not in existence during the years when the radiation data was collected in a series of scientific studies by Brock and by Cech and coinvestigators.

The Vital Records data on birth defects and the Birth Defect Epidemiology and Surveillance Branch Registry data are not identical. The Registry utilizes active surveillance by trained workers, who physically visit delivery facilities and abstract medical records. Marengo (71) calculated the sensitivity of the Bureau of Vital Statistics (BVS) data in recording selected birth defects, using the Texas Birth Defect Registry as the standard. For 1999 to 2000, the BVS records on birth defects were less accurate than the registry records. However, for cleft lip and palate, the concordance was 86% of the time. This was better than for other birth defects.

The present study shares some of the limitations typical of retrospective ecological studies. There were no direct radiation measurements of tap water in the case households, but the measurements were extrapolated from the data we obtained for the distribution systems. In addition, residential migration patterns within the study area, or into and out of the study area, were not evaluated. We believe that misclassification error by residence at birth versus conception would have been distributed evenly throughout the study area and not have affected the relative differences between the areas of lower versus higher concentrations of radionuclides. Still further, there were no known prior clusters of genetic syndromes that may contribute to elevated rates of orofacial cleft defects in northwest Harris County. While genetic clusters were beyond the scope of this study, it is difficult to envision and identify such distinct genetic clusters in a cosmopolitan population such as that of the Greater Houston/Harris County, with a multiethnic and multiracial population. However, multivariate analyses indicated that race/ethnicity was not a significant factor associated with cleft birth rates.

With these limitations, the existence of historical data on Ra-226 and Rn-222 was a major strength, and considered to overcome the compromise of using older BVS records for cleft births for 1990 to 1994. It is our opinion that the underreporting on cleft births in BVS records would have been distributed uniformly within Harris County and would minimally change the relative differences in rates within the study area.

The radiation issues continue to exist in northwest Harris County, (72) but the data quality reported by the public utilities for years matching the birth defect registry record (1999-2002) was not an improvement over the data first reported in Harris County. The Houston/Harris County monitoring and reporting system was criticized in a 2003 by the Natural Resources Defense Council. (73) Until 2003, the test protocol and regulatory compliance limits for radionuclides in drinking water were determined by samples taken from a vaguely defined "representative point to the water distribution system." (74) Under such a protocol, some public utilities may not have exceeded the regulatory limits for radionuclides. New EPA rules (promulgated in 2003) require individual water well entry points to the distribution system to meet the radionuclide limits.

Presently, we are expanding our investigation in collaboration with colleagues from the Birth Defect Epidemiology and Surveillance Branch Registry. The advantage of using these more recent records is tempered by the fact that Rn-222 data are not available for these recent years. While with some caveats, Ra-226 data could be ascertained from the public utility records, Rn-222 is not monitored under the Safe Drinking Water Act and thus, data are not being gathered by any regulatory agencies. New household-level data on radon would need to be generated, which is contingent upon the availability of funding and significant technical resources.

Geospatial patterns alone are not sufficient evidence that elevated levels of radioactivity in the drinking water are causally related to the cleft defect births in the same area of Harris County, Texas. Genetic and other environmental risk factors are too complex and intertwined to be disentangled simply by comparing birth defect and environmental patterns. A more rigorous analytical treatment of the subject is necessary, such as with a concurrent prospectively matched case-control study design. (75) Clarification is needed whether radium and radon exposures are direct risk factors for cleft defects, or if they are sentinels for other environmental problems related to previous land uses, such as hydrocarbon extraction, pesticides, and herbicide use in the same area. To address these issues, we have begun a pilot incident case-control investigation looking for biomarkers of exposure to uranium 238 and selected organic chemicals (pesticides and herbicides).

In spite of limitations, the association that we observed between elevated concentrations of uranium decay products and elevated orofacial cleft birth anomalies in Harris County, Texas, deserves attention. This association was (a) consistent with reports elsewhere, (42) (b) biologically plausible, (37) (c) temporally consistent, such that potential exposure preceded birth defect manifestation (late 1980s and early 1990s), (d) consistent in terms of geospatial distribution (northwest Harris County), and (e) defect specific. Taking this into consideration, even though the risk from other agents may markedly differ geographically, it would be unwise to ignore the possible contribution of a water-conveyed factor.

Acknowledgments

The authors are grateful to Mr. John Marostica of the Bureau of Vital Statistics of Texas Department of Health for providing birth certificate data.

References

1. World Health Organization. ICD-9-CM: International Classfication of Diseases. Geneva, Switzerland: WHO; 2006.

2. Coleman JR Jr, Sykes JM. The embryology, classification, epidemiology, and genetics of facial clefting. Facial Plast Surg Clin North Am 2001;9:1-13.

3. James WH. Are oral clefts a consequence of maternal hormone imbalance? Evidence from the sex ratios of sibs of probands. Teratology 2000;62:342-345.

4. Felix TM, Hanshaw BC, Mueller R, et al. CHD7 gene and non-syndromic cleft lip and palate. Am J Med Gene A 2006;140:2110-2114.

5. Jakobsen LP, Knudsen MA, Lespinasse J, et al. The genetic basis of the Pierre Robin sequence. Cleft Palate Craniofacial J 2006;43:155-159.

6. Nie X, Luukko K, Kettunen P. FGF signaling in craniofacial development and developmental disorders. Oral Dis 2006;12:102-111.

7. Prescott NJ, Winter RM, Malcolm S. Maternal MTHFR genotype contributes to the risk of non-syndromic cleft lip and palate. J Med Genet 2002;39:368-369.

8. Fallin MD, Hetmanski JB, Park J, et al. Family-based analysis of MSX1 haplotypes for association with oral clefts. Genet Epidemiol 2003;25:168-175.

9. Murray JC, Schutte BC. Cleft palate: players, pathways, and pursuits. J Clin Invest 2004;113:1676-1678.

10. Owens JR, Jones JW, Harris F. Epidemiology of facial clefting. Arch Dis Child 1985;60:521-524.

11. Wyszynski DF, Nambisan M, Surve T, et al. Increased rate of major malformations in offspring exposed to valproate during pregnancy. Neurology 2005;64:961-965.

12. Medina-Carrilo L, Rivas-Solis F, Fernandez-Arguelles R. Risk for congenital malformations in pregnant women exposed to pesticides in the state of Nayarit, Mexico. Ginecol Obstet Mex 2002;70:538-544.

13. Chevrier C, Dananche B, Bahuau M, et al. Occupational exposure to organic solvent mixtures during pregnancy and the risk of nonsyndromic clefts. Occup Environ Med. Available at: http://oem.bmjjournals.com/cgi/content/abstract/oem.2005.024067vl, 2006. Accessed March 31, 2007.

14. Prescott NJ, Winter RM, Malcolm S. Nonsyndromic cleft lip and palate: complex genetics and environmental effects. Ann Hum Genet 2001;65:505-515.

15. Murray JC. Gene/environment causes of cleft lip and/or palate. Clin Genetics 2002;61:248-256.

16. Shaw GM, Wasserman CR, Lammer EJ, et al. Orofacial clefts, parental cigarette smoking, and transforming growth factor-alpha gene variants. Am J Hum Genet 1996;58:551-561.

17. Lammer EJ, Shaw GM, Iovannisci DM, et al. Maternal smoking and the risk of orofacial clefts: susceptibility with NAT 1 and NAT 2 polymorphisms. Epidemiology 2004;15:150-156.

18. van Rooij IA, Wegerif MJ, Roelofs HM, et al. Smoking, genetic polymorphisms in bio-transformation enzymes, and nonsyndromic oral clefting: a gene-environment Interaction. Epidemiology 2001;12:502-507.

19. Zeiger JS, Beaty TH, Liang KY. Oral clefts, maternal smoking, and TGFA: a meta-analysis of gene-environment interaction. Cleft Palate Craniofac J 2005;42:58-63.

20. Hwang SJ, Beaty TH, Panny SR, et al. Association study of transforming growth factor alpha (TGF alpha) TaqI polymorphism and oral clefts: indication of gene-environment interaction in a population-based sample of infants with birth defects. Am J Epidemiol 1995;141:629-636.

21. Christensen K, Olsen J, Norgaard-Pedersen B, et al. Oral clefts, transforming growth factor alpha, gene variants, and maternal smoking: a population-based case-control study in Denmark, 1990-1994. Am J Epidemiol 1999;149:248-255.

22. Krapels IP, van Rooij IA, Ocke MC, et al. Maternal dietary B vitamin intake, other than folate, and the association with orofacial cleft in the offspring. Eur J Nutr 2004;43:7-14.

23. van Rooij IA, Swinkel DN, Bloom HJ, et al. Vitamin and homocysteine status of mothers and infants and risk of nonsyndromic orofacial clefts. Am J Obstet Gynecol 2003;189:1155-1160.

24. van Rooij IA, Groenen PM, van Drongelen M, et al. Orofacial cleft and spina bifida: N-acetylthransferase phenotype, maternal smoking, and medication use. Theratology 2002;66:260-266.

25. Prescott NJ, Malcolm S. Folate and the face: evaluating evidence for the influence of folate genes on craniofacial development. Cleft Palate Craniofacial J 2002;39:327-331.

26. Womersley J, Stone DH. Epidemiology of facial clefts. Arch Dis Child 1987;62:717-720.

27. Brent RL, Beckman DA. The contribution of environmental teratogens to embryonic and fetal loss. Clin Obstet Gynecol 1994;37:646-670.

28. Nomura T. X-ray-and chemically induced germ-line mutation causing phenotypical anomalies in mice. Mutat Res 1988;198:309-320.

29. Brent RL. Radiation teratogenesis teratogen update. In: Brent RL, editor. Environmentally Induced Birth Defect Risks. New York, New York: Alan R Liss; 1986; pp. 145-163.

30. Kallen B, Karlsson P, Lundell M, et al. Outcome of reproduction in women irradiated for skin hemangioma in infancy. Radiat Res 1998;149:202-208.

31. Ritenour ER. Health effects of low level radiation: carcinogenesis, teratogenesis, and mutagenesis. Semin Nucl Med 1986;16:106-117.

32. Zieglowski V, Hemprich A. Facial cleft birth rate in former East Germany before and after the reactor accident in Chernobyl. Mund Kiefer Gesichtschir 1999;3:195-199.

33. Yamazaki JN, Wright SW, Wright PM. A study of the outcome of pregnancy in women exposed to the atomic bomb blast in Nagasaki. J Cell Physiol Suppl 1954;43:319-328.

34. Sulik KK, Cook CS, Webster WS. Teratogens and craniofacial malformations: relationships to cell death. Development 1988;103:213-231.

35. Scherb H, Weigelt E. Cleft lip and cleft palate birth rate in Bavaria before and after the Chernobyl nuclear power plant accident. Mund Kiefer Gesichtschir 2004;8:106-110.

36. Rugh R. Effects of Ionizing Radiations, Including Radioisotopes, on the Placenta and Embryo. Birth Defects Original Article Series. New York: The National Foundation-March of Dimes; 1965.

37. Purnell SJ, Allen JE, Oyedepo AC, et al. Fetal dosimetry from natural alpha-particle emitters. Radiat Res 1999;152:S133-S136.

38. Persaud TVN. Environmental Causes of Human Birth Defects. Springfield, IL: Charles C. Thomas; 1990.

39. Kendall GM, Smith TJ. Doses from radon and its decay products to children. J Radiol Prot 2005;25:241-256.

40. Joss R, Brunner K. Pregnancy following tumor therapy. Schweiz Med Wochenschr 1979;109:904-907.

41. Risk estimation for multifactorial diseases: a report of the International Commission on Radiological Protection. Ann ICRP 1999;29:1-144.

42. Gentry JT, Parkhurst E, Bulin GV Jr. An epidemiological study of congenital malformations in New York State. Am J Public Health 1959;49:497-513.

43. Centers for Disease Control and Prevention. Second National Report on Human Exposure to Environmental Chemicals: Uranium. In: NCEH Pub. No. 02-0716. Atlanta, GA; 2003; pp. 23-25.

44. Shoji S, Watanabe H, Katoh O, et al. Developmental malformations and intrauterine deaths in gamma-ray-irradiated scid mouse embryos. Int J Radiat Biol 1998;73:705-709.

45. Beebe GW. Ionizing radiation and health. Am Sci 1982;70:35-44.

46. Bladen CL, Dynan WS, Kozlowski DJ. Low dose radiation damage and radioprotection in the vertebrate embryo. Office of Biological and Environmental Research, DOE Low-dose Radiation Program Workshop IV. Available at: http://www.lowdose.energy.gov.Accessed March 31, 2007.

47. Brent RL, Gorson RO. Radiation exposure during pregnancy. Curr Probl Diagn Radio 1972;2:1-48.

48. Arfsten DP, Still KR, Ritchie GD. A review of the effects of uranium and depleted uranium exposure on reproduction and fetal development. Toxicol Ind Health 2001;17:180-191.

49. Brock JD. Radioactivity of drinking water in northwest Harris County. Houston, Texas: Master Thesis, University of Texas Health Science Center at Houston School of Public Health; 1984.

50. Cech I. Assessment of radium and radon in water. Geol Soc Am Bull 1987;29:8-9.

51. Cech I, Lemma M, Prichard HM, et al. Radium-226 and radon-222 in domestic water of Houston-Harris County, Texas. In: Barbara G, editor. Radon, Radium, and Other Radioactivity in Ground Water: Hydrologic Impact and Application to Indoor Airborne Contamination, Proceedings of the National Water Well Association Conference in Somerset, New Jersey. Chelsea, MI : Lewis Publishers; 1987; pp. 377-402.

52. Cech I, Kreitler CH, Prichard H, et al. Radon distribution in domestic water of Texas. Ground Water 1988;26:561-569.

53. Cech I, Lemma M, Kreitler CW, et al. Radium and radon in water supplies from the Texas gulf coastal aquifer. Water Res 1988;22:109-121.

54. Cech I, Prichard HM, Mayerson A, et al. Pattern of distribution of radium 226 in drinking water of Texas. Water Resources Research 1987;23:1987-1995.

55. Cech I, Siegle P, Merritt M. Toward a better understanding of the geographic distribution of radon in domestic water supplies. In: Pawlowski L, editor. Proc of Conference Chemistry for the Protection of the Environment. Lublin, Poland: United States Department of Energy/United States Environmental Protection Agency; 1992; pp. 201-213.

56. United States Environmental Protection Agency. National Primary Drinking Water Regulations; Radionuclides; Final Rule. In: Vol 65: Federal Register. December 7, 2000; pp. 76707-76753. Available at: http://www.epa.gov/safewater/rads/radfr.html. Accessed March 31, 2007.

57. MapInfo Corporation. StreetInfo, Texas West, Enhanced Address Layer, version 7.0. Troy, NY; 2003.

58. MapInfo Corporation. MapInfo, Version 8. Troy, NJ; 2004.

59. Kuldorff M. Information Management Services I. SaTScan v 3.0.5. Software for the spatial and space-time scan statistics. Bethesda, MD: National Cancer Institute; 2002.

60. United States Environmental Protection Agency. Proposed Radon in Drinking Water Rule. In: 64 Federal Register 59246, 59328; November 2, 1999.

61. Spritz RA. The genetics and epigenetics of orofacial clefts. Curr Opin Pediatr 2001;13:556-560.

62. Katy Area Chamber of Commerce. Katy Business, Katy Community and Relocation to Katy, Texas; 2003.

63. Burg JAR. Letter to residents. Atlanta, Georgia: Department of Health and Human Services, Agency for Toxic Substances and Disease Registry; 1991.

64. Burg JAR. Letter to physicians. Atlanta, Georgia: Department of Health and Human Services Agency for Toxic Substances and Disease Registry 1991.

65. Dixon R. Toxic response of reproductive system. In: Doull J, editor. Casarett and Doull's Toxicology: The Basic Science of Poisons. New York: McGraw-Hill, Health Professions Division; 1995; pp. 432-477.

66. Miller RW. Measures of reproductive effects. Chemical and radiation hazards to children. In: Report of the 84th Ross Conference on Pediatric Research. Columbus, OH: Ross Laboratories; 1982; pp. 85-89.

67. Polednak AP, Stehney AF, Rowland RE. Mortality among women first employed before 1930 in the United States radium dial-painting industry: a group ascertained from employment lists. Am J Epidemiol 1978;107:179-195.

68. Mays CW, Spiess H. Bone tumors in thorotrast patients. Environ Res 1979;18:88-93.

69. Mays CW, Spiess H. Bone sarcoma in patients given radium-224. In: Boice JD, Fraumeni JF Jr (eds). Radiation Carcinogenesis, Epidemiology and Biological Significance. New York, Raven Press, 1984, pp 241-252.

70. Mays CW, Van Dilla MA, Floyd RL, et al. Radon retention in radium-injected beagles. Radiat Res 1958;8:480-489.

71. Marengo L. Comparison of Texas Birth Defect Registry and Texas Vital Record Data for Selected Birth Defects Readily Diagnosed at Delivery. Austin, TX: Texas Birth Defect Epidemiology and Surveillance Branch, Department of State Health Services; 2004.

72. Texas Commission on Environmental Quality. Drinking water quality records. Austin, TX; 2006.

73. Olson E. What's on Tap? Grading Drinking Water in US Cities. New York, NY: Natural Resources Defense Council. Available at: http://www.nrdc.org/water/drinking/uscities/pdf/whatsonrtap.pdf. Accessed March 31, 2007.

74. Harris County Municipal Utility District 261. Facts surrounding recent developments. Letter to all district constituents. June 11, 2004.

75. Cech I, Burau K, Al-Hashimi R, et al. Chemical and Biological Risk Factors in the Etiology of Neural Tube Birth Defects Near the U.S.-Mexico Border, Final Report to CDC/ATSDR, December 2006.

How far can you go without destroying from within what you are trying to
defend from without?
--Dwight Eisenhower

Irina Cech, PhD, Keith D. Burau, PhD, and Jane Walston, BS, MPH

From the University of Texas at Houston, Health Science Center, School of Public Health, Houston, TX.

This study was conducted with the approval of the Committee for the Protection of Human Subjects, University of Texas Health Science Center at Houston.

Reprint requests to Dr. Irina Cech, The University of Texas at Houston Health Science Center School of Public Health, RAS E1007, 1200 Herman Pressler, Houston, TX 77030. Email: [e-mail]

The authors have no commercial or proprietary interests in any of the methods, equipment, or computer software used for analysis of water samples.

Accepted October 17, 2006.

RELATED ARTICLE: Key Points

* Rapid urban expansion involves the risk of encroachment into areas where previous land uses might not have been fully compatible with the development of residential water wells.

* Attention of physicians is invited to radium and radon as unsuspected and unnecessary health risk factors in some tap water.

* A cluster of significantly greater numbers of cleft defect births relative to the total numbers of live births has been documented in the northwest part of Harris County, Texas, where elevated levels of radium and radon have been known since the 1980s.

* Geographic areas with a high probability of exposure to elevated levels of domestic radiation need to be targeted for studies of cleft and other cranioskeletal birth defects.

Table 1. Mean annual incidence rates of cleft birth defects by radium-
226 concentrations in the domestic water; univariate and adjusted for
potentially confounding variables, Harris County, Texas, 1990-1994

Descriptive statistics

Dependent variable: mean annual rate

Radium  Mean     Standard deviation  N

1        5.4113   6.45735            124
2       14.5600  12.47933             10
Total    6.0940   7.41108            134

1 = Ra-226 <3 pCi/L; 2 = Ra-226 [greater than or equal to]3 pCi/L.

Tests of between zip code effects

Dependent variable: mean annual rate

                 Type III sum of
Source           squares          df   Mean square  F       Sig.

Corrected model    774.527 (a)      1   774.527     15.656  0.000
Intercept         3690.873          1  3690.873     74.604  0.000
RADIUM             774.527          1   774.527     15.656  0.000
Error             6530.388        132    49.473
Total            12281.300        134
Corrected total   7304.915        133

(a) R squared = 0.106 (adjusted R squared = 0.099).

Tests of between zip code effects

Dependent variable: mean annual rate

                 Type III sum of
Source           squares          df   Mean square  F       Sig.

Corrected model   1523.736 (a)      8  190.467       4.118  0.000
Intercept           94.325          1   94.325       2.039  0.156
OCCUPATI           154.463          1  154.463       3.340  0.070
PER_CAPI            63.719          1   63.719       1.378  0.243
EDUCATIO            67.407          1   67.407       1.457  0.230
PERC_WHI            88.465          1   88.465       1.913  0.169
PERC_BLA            93.472          1   93.472       2.021  0.158
PERC_SPA            94.116          1   94.116       2.035  0.156
PERC_ASI            82.208          1   82.208       1.777  0.185
RADIUM             611.236          1  611.236      13.216  0.000
Error             5781.180        125   46.249
Total            12281.300        134
Corrected total   7304.915        133

(a) R squared = 0.209 (adjusted R squared = 0.158).
Where covariables were: Occupation = ratio of white to blue collar
workers per zip code; per capita income, U. S. dollars; education,
percent <9 years of school completed; percent white population; percent
black population; percent Hispanic population; percent Asian population.

Table 2. Mean annual incidence rates of cleft birth defects by radon-222
concentrations in domestic water; univariate and adjusted for
potentially confounding variables, Harris County, Texas, 1990-1994

Descriptive statistics

Dependent variable: mean annual rate

Radon  Mean    Standard deviation  N

1      5.4670   6.41436            112
2      9.2864  10.84454             22
Total  6.0940   7.41108            134

1 = Rn-222 <300 pCi/L; 2 = Rn-222 [greater than or equal to]300 pCi/L.

Tests of between zip code effects

Dependent variable: mean annual rate

                 Type III sum of
Source           squares          df   Mean square  F       Sig.

Corrected model    268.242 (a)      1   268.242      5.032  0.027
Intercept         4002.358          1  4002.358     75.080  0.000
RADON              268.242          1   268.242      5.032  0.027
Error             7036.674        132    53.308
Total            12281.300        134
Corrected total   7304.915        133

(a) R squared = 0.037 (adjusted R squared = 0.029).

Tests of between zip code effects

Dependent variable: mean annual rate

                 Type III sum of
Source           squares          df   Mean square  F      Sig.

Corrected model   1126.674 (a)      8  140.834      2.849  0.006
Intercept           94.031          1   94.031      1.902  0.170
OCCUPATI            74.214          1   74.214      1.502  0.223
PER_CAPI             0.339          1    0.339      0.007  0.934
EDUCATIO            59.059          1   59.059      1.195  0.276
PERC_WHI            88.081          1   88.081      1.782  0.184
PERC_BLA            94.176          1   94.176      1.905  0.170
PERC_SPA            94.725          1   94.725      1.917  0.169
PERC_ASI            79.023          1   79.023      1.599  0.208
RADON              214.174          1  214.174      4.333  0.039
Error             6178.241        125   49.426
Total            12281.300        134
Corrected total   7304.915        133

(a) R squared = 0.154 (adjusted R squared = 0.100).
Where covariables were: Occupation = ratio of white to blue collar
workers per zip code; per capita income, U. S. dollars; education,
percent <9 years of school completed; percent white population; percent
black population; percent Hispanic population; percent Asian population.