Review of Louisiana Tumor Registry Cancer Incidence Data
AGRICULTURE STREET LANDFILL
NEW ORLEANS, ORLEANS PARISH, LOUISIANA
As a follow-up activity to recommendations outlined in a prior health consultation regarding theAgriculture Street Landfill Superfund (ASL) site in New Orleans, the Louisiana Department ofHealth and Hospitals/Office of Public Health/Section of Environmental Epidemiology andToxicology (LDHH/OPH/SEET) reviewed available health outcome data on additional years ofcancer incidence data for the Agriculture Street Landfill community. These additional yearsinclude 1994-1997, and are the most current data available. Prior health consultations reviewedhealth outcome data from 1983-1993. The health outcome data examined in this healthconsultation include cancer incidence rates (from the Louisiana Tumor Registry (LTR)). Thishealth consultation serves to provide the residents with a way to compare the health effects intheir area with a similar, but larger population. No cause-and-effect relationship with site related contaminants can be determined by this type of review.
The ASL site is a 95-acre site located in New Orleans, Orleans Parish, Louisiana. The site wasused as a municipal landfill receiving municipal waste and construction debris for more than 50years prior to being developed for housing and businesses. The landfill was closed in 1965. During the 1970s and 1980s, Gordon Plaza Subdivision, Housing Authority of New Orleans(HANO) residences, Gordon Plaza Apartments, the Moton School, the Press Park residential areaand community center were constructed over part of the landfill. Forty-eight acres of the landfillremain undeveloped and fenced. Metals, pesticides, and polycyclic aromatic hydrocarbons(PAHs) have been found in surface and subsurface soils throughout the site during environmentalstudies.
In December of 1994, the U.S. Environmental Protection Agency (EPA) placed the ASL site onthe National Priorities List (also called the Superfund list). EPA conducted a Remedial/RemovalIntegrated Investigation (RRII) of the entire site and released their results in 1995. During thisinvestigation, surface soil, subsurface soil, groundwater, tap water, air, and indoor dust sampleswere analyzed for chemicals found at the site . Based on those environmental results andhealth data, OPH/SEET in conjunction with the Agency for Toxic Substances and DiseaseRegistry (ATSDR), released a Public Health Assessment.
The conclusions of the Public Health Assessment were:
- The undeveloped area of the site was classified as a public health hazard. The highestlevels of contaminants have been found in the undeveloped area. Although entry to thisarea has been limited by a fence, individuals continue to access this area and may come incontact with the elevated levels of lead and arsenic in the soil. If this area was developedfor future residential use as is, exposure to lead, arsenic, and polycyclic aromatichydrocarbons (PAHs) in the soil could pose an unacceptable health risk to residents.
- The majority of the residential area and the Press Park Community Center has beenclassified as no apparent public health hazard since the levels of contaminants in thesoil are generally below levels that may cause health problems. There are scatteredpockets of lead, arsenic, and PAHs in soil that need to be addressed to limit anypossibility of exposure to levels of health concern.
- The contamination presented at the Moton School poses no public health hazard sincethe levels of chemicals in the soil, air, and water were well below levels that may causehealth problems.
The remedy for the site is completed and no further action is planned. During the removalaction, two feet of soil was removed prior to grading and a semi-permeable geotextile liner wasinstalled and clean soil was backfilled. The first five-year review is ongoing. The integrity of thecap was inspected by EPA and its contractor, and the Louisiana Department of EnvironmentalQuality on October 17, 2002. Permanent relocation is an issue which continues to polarize thecommunity since many prefer redevelopment. The Superfund law was examined by EPA whoconcluded that relocation is not possible because the redevelopment remedy is attainable andprotective.
Previous health outcome investigations concerning the ASL site include a review of blood leadlevels and a health survey conducted by Xavier University. Both were addressed in the 1996Public Health Assessment . The blood lead data were evaluated because lead is one of themost prevalent contaminants at the site and children are especially sensitive to the toxic effects oflead. The results of the 1993 and 1994 blood lead screens, conducted by the city of New Orleans,indicated that the children tested who live on the ASL site had a lower percentage of elevatedblood leads levels as compared to other children in New Orleans (18% and 44%, respectively).
The Deep South Center for Environmental Justice at Xavier University of Louisiana conducted ahealth survey of 328 adults and children (approximately one third of the residents of ASL) wholive in Gordon Plaza and Press Park . The survey did not include a comparison group ofindividuals living outside the ASL area. Even though the findings indicated that there arereported illnesses in the community, the lack of a comparison group makes it impossible todetermine if the illnesses at the ASL site are occurring more often than is expected for acommunity not living on a Superfund site.
The Office of Public Health has taken a very active role in the activities occurring at the ASLsite. Initially, Louisiana Tumor Registry (LTR) data was examined in 1997. Also examined atthis time, were birth weights of children born in the area. A follow-up examination of thisinformation occurred in 2001. Results of both these examinations showed similar elevations inbreast cancer incidence for the area that includes the ASL site. To follow-up to those results, weare once again examining cancer incidence for the most recent time period for which it exists.
Health outcome data reviewed for a site are generally based upon (1) possible health effects thatcould be caused by exposure to site contaminants and (2) the availability of data. That is why inthe initial Public Health Assessment, information on blood lead levels of the ASL children wasreviewed. Lead has been a contaminant of concern at the ASL site and blood lead data wereavailable.
For this report, OPH reviewed additional years of cancer incidence not included in the priorreview. The period of time selected for this recent evaluation of the cancer incidence data was 1994-1997, which is the most recent health data available. The smallest geographic area for which we can calculate rates of disease is the census tract.
In order to compare the ratios of cancer incidence around the ASL site with parish or regionalrates, it is necessary to have specific population data. Population data, categorized by age, andhealth outcome data are both available at the census tract level. Census tracts are subdivisions ofparishes. They usually have between 2,500 and 8,000 persons and are designed to be relativelyhomogeneous or similar with respect to population characteristics, economic status, and livingconditions .
The ASL site lies within Census Tract 1703 of Orleans Parish and covers about 1/5 of the censustract area. The total population for Census Tract 1703 is 4,506 persons, according to the 1990census data . The population for the ASL site was estimated to be 1,137 persons, based on thenumber of housing units . Thus, the population on the ASL site is estimated to be about 1/4 ofthe census tract population.
Table 1 summarizes the 1990 Census information for Louisiana, Orleans Parish, and CensusTract 1703, which contains the ASL site. Review of the census data suggests that Orleans Parishand Census Tract 1703 have a higher percentage of African Americans than the state as a whole.Results occurring in the white demographic strata would need to be examined closely as thenumbers comprising this strata are small, and small number sometimes yield unstable results.Orleans Parish and Census Tract 1703 also have a higher percentage of persons and familiesliving below the poverty level. Census Tract 1703 has a lower per capita income than Louisianaand Orleans Parish. Median family and household incomes are also lower for the census tract than Orleans Parish and Louisiana.
|1990 CENSUS DATA|
|LOUISIANA||ORLEANS PARISH||CENSUS TRACT 1703 ASL|
|All Persons |
|<7 Years |
|476,687 (11.3) |
| 54,365 (10.9) |
|510 (11.3) |
|Number of Families||1,098,374||119,516||1,062|
|Number of Households||1,498,371||187,662||1,543|
|Median Family |
|Persons Below |
|Median Year Housing Built||1969||1951||1963|
1. Method for analyzing cancer incidence data
Cancer incidence data were obtained for the 10-year period of 1988-1997 from the LouisianaTumor Registry. These are the most recent cancer data that are available by census tract levels. Because the community has expressed concern about cancer in general and because the currentlevel of contaminants would not be expected to cause an increase in any particular cancer, OPHreviewed all types of cancers that were diagnosed in the census tract. Cancer incidence(occurrence of cancer) was chosen for this review because cancer mortality (death) rates areaffected by how advanced the cancer was at the time of diagnosis, access to health care, and otherfactors not related to exposure.
In order to evaluate whether the Census Tract 1703, which contains the ASL site, has an elevatedlevel of cancer incidence, the region was chosen as a comparison population. In general, thecomparison population should be large enough so that its cancer rates are stable (that is, the ratesdo not fluctuate greatly). Furthermore, the comparison population should be similar to thepopulation being studied in factors which could affect disease rates, such as socioeconomicfactors and racial distribution, other than the study exposure factors of interest. Therefore, theLouisiana Tumor Registry's Region I, which includes Jefferson, Orleans, and St. BernardParishes, was chosen as the comparison population.
Because different groups of people have different rates of cancer, the cancer incidence data werecalculated separately by age group, sex, and race. For example, as we get older, our chance ofgetting cancer increases. Therefore, an older population would be expected to have a higher rateof cancer than a younger one. In making our comparisons, the number of cancers for a certainage group is compared to the number of cancers expected for that same age group.
Once the rates are calculated for specific age groups, sexes, and races, the standardized incidenceratios (SIRs) are calculated. The SIR estimates the occurrence of cancer in the study population(in this case, Census Tract 1703) relative to what might be expected if the census tract had thesame cancer rate as the comparison population (Region I). An SIR is the ratio of the observednumber of cases to the expected number of cases.
The SIR tells us how much higher or lower the census tract's cancer rate is compared to that inthe other population. If the observed number of cases equals the expected number of cases, theSIR will equal 1. If there are more observed cases than one would expect, then the SIR will begreater than 1. If there are less observed cases than one would expect, then the SIR will be lessthan 1. For example, if 10 cases are observed in the study population, but five cases wereexpected, then the SIR = 10/5 = 2, and the area has two times the cancer rate than expected. Butif 20 cases were expected, then the SIR = 10/20 = 0.5, and the area has half the rate thanexpected.
Caution should be exercised, however, when interpreting the SIR. The interpretation must takeinto account the actual number of cases observed and expected, not just the ratio. Two SIRs canhave the same number, but represent very different scenarios. For example, a SIR of 1.5 couldmean three cases were observed and two were expected (3/2 = 1.5). Or it could mean 300 caseswere observed and 200 were expected (300/200 = 1.5). In the first instance, only one excesscancer occurred, which could easily have been due to chance. But, in the second instance, 100excess cancers occurred and it would be less likely that this would occur by chance alone.
To help interpret the SIR, the statistical significance of the difference can be calculated. In other words, the number of observed cases can be determined to be significantly different from the expected number of cases or the difference can be due to chance alone. "Statistical significance" for this review means that there is less than five percent chance (p-value <0.05) that the observed difference is merely the result of random fluctuation in the number of observed cancer cases. If the SIR is found to be statistically significant, then the difference between the expected and observed cases is probably due to some set of factors that influences the rate of that disease. Because cancer is, unfortunately, so common (more than one in three of us will develop cancer in our lifetime), every community will experience a certain number of cancers. Through the years, you would expect some fluctuation in the numbers. One year, there may be a few more cases of cancer A and the next year a few less. This occurs by chance. There is no specific cause. Just like flipping a coin, although you expect that you will get heads half the time and tails half the time, it doesn't always come up even. Out of 10 coin tosses, you may get seven heads and three tails or four heads and six tails. The more tosses you make, the closer you will probably come to getting a 50-50 mix. This is why, in order to determine if cancer rates are elevated, the statistical significance must be considered.
2. Standardized Incidence Ratios for Census Tract 1703 for the years, 1988-1997.
For Census Tract 1703, all major groupings of cancer were evaluated: esophageal, stomach,colon, rectal, liver, other biliary, pancreatic, lung, soft tissue, brain, breast, cervix uteri, corpusuteri, ovarian, prostate, bladder, kidney, thyroid, Hodgkin's lymphoma, Non-Hodgkin'slymphoma, multiple myeloma, leukemia, and all cancers combined. Table 2 shows the numberof cancers that were observed in the census tract for the 10-year period of 1988-97. A prior healthconsultation reviewed cancer incidence from 1988-1993, however, this was combined with thenew data, to produce more stable results.
Standardized incidence ratios (SIRs) were calculated for each type of cancer when five cases ormore were observed in the census tract in the 10-year period. Calculating SIRs with fewer casesleads to statistical instability. The New Orleans Region (Jefferson, Orleans, & St. BernardParishes) was used as the comparison population. The census tract calculations were based onthe 1990 population data and the regional rates are also based on 1990 census data plus estimates from years between census surveys.
|All, white males||18||7.09||2.54||Yes|
|All, white females||10||4.20||2.38||Yes|
|All, black males||60||67.60||0.89||No|
|All, black females||75||62.73||1.20||No|
|Colon, black females||5||6.54||0.76||No|
|Lung & bronchus||28||30.37||0.92||No|
|Lung & bronchus, whites||5||2.26||2.21||No|
|Lung & bronchus, blacks||23||25.14||0.92||No|
|Lung & bronchus, males||16||18.70||0.86||No|
|Lung & bronchus, females||12||11.62||1.03||No|
|Lung & bronchus, black males||13||16.77||0.77||No|
|Lung & bronchus, black females||10||8.36||1.20||No|
|Breast, white females||6||1.34||4.47||Yes|
|Breast, black females||30||18.86||1.59||Yes|
|Prostate (males only)||24||16.47||1.46||No|
|Prostate (males only), black males||21||16.80||1.25||No|
|Other, Ill-defined & Unknown||7||3.82||1.83||No|
|Other, Ill-defined & Unknown, blacks||6||4.21||1.42||No|
From Table 2, one can see that the observed number of lung, colon, prostate, and total cancerssites combined are not statistically different than what was expected. However, there was astatistically significant excess of breast cancer in Census Tract 1703 from the years 1988-1997. This excess was seen in white males and white females.
3. Breast Cancers for Census Tract 1703 for 1983-87 and 1988-97
Adding the 1983-87 data with the 1988-97 data provides an 15-year span of breast cancerincidence in this census tract. The population used for the census tract calculations was based onthe estimated 1985 and the 1990 census data. For this time period, 1983-97, no statisticallysignificant excess was seen for all females combined or black females. However, a statisticallysignificant excess of breast cancer did occur in white women (three cases expected and 10 casesobserved) in Census Tract 1703. Table 3 is a summary of the observed breast cancer cases, the expected cases, and the SIRs for the three time periods.
|All Females||Black Females||White Females|
* Statistically significant at p=0.05 level.
4. Discussion of Excess Cancers
Observations about the breast cancer rates for Census Tract 1703 are summarized as follows: Astatistically significant increase for all women is seen in the 1988-1997 period, but is notobserved for the entire 15-year period, 1983-97. White women in this census tract do not show astatistically significant elevation in breast cancer for either 1983-87, but show a statisticallysignificant elevation in breast cancer for the 1988-97 period. The SIRs for all three periods wereelevated. Small numbers makes it more difficult to achieve statistical significance. When thesmall numbers are combined, the 15 year rate is statistically significant.
The review of cancer incidence data in this document is only a screening mechanism to alert usto unusual rates. No cause or reason for the excess cancer in Census Tract 1703 can bedetermined at this point. A possible factor in the breast cancer elevation in white women is thelow percentage of whites in the area coupled with the small numbers of cancer.
A number of factors are known to be associated with a higher risk of developing breast cancer. These include: older age, family history of breast cancer, early menstruation, late menopause,recent use of oral birth control pills, never having children, or having your first child at a late age. Other non-reproductive factors include radiation exposure, consumption of dietary fat, and bodysize. In addition to these, physical activity, alcohol consumption, and pesticide and chemicalexposure are suspected risk factors, and are currently being studied to determine their impact on breast cancer risk [1,2,3]. No survey has been conducted in this census tract of the above riskfactors. Therefore, we do not know if any of these have influenced the excess rate of breastcancers seen in Census Tract 1703 for some of the population groups.
Although some environmental contaminants have the potential to affect breast cancer risk, clearenvironmental links to breast cancer are limited, with the exception of radiation exposure andalcohol intake. Some studies have suggested links with certain pesticides, such as DDT. However, more research is needed to establish these chemicals as possible risk factors for breastcancer .
The current environmental data for the ASL site, that was collected by EPA, did not showpesticides in the soil at levels that are known to cause health effects. In fact, the EPA RRII reportstates that for surface soils, "pesticides found on site are not significantly different in type andconcentration than those found in the background" . Other chemicals found at the ASL site,such as metals and PAHs, have not been associated with breast cancer. In addition, removal andremedial activity at the site has further reduced the amount of site contaminants.
It is extremely difficult to identify the cause or causes of elevated rates of a chronic disease, suchas breast cancer, especially in a small population. Since cancers may take many years to develop,various genetic, lifestyle, and environmental factors may interact before the disease becomesapparent. Also, it may be difficult to clearly identify those risk factors when they occurred yearsearlier. Another problem is that a factor may be related differently to the initial development of the disease than to its later course .
LDHH/OPH/SEET reviewed cancer incidence data from 1983-1997. The following conclusionscan be made based on the data reviewed:
- The census data suggests that the Census Tract 1703 (which includes the ASL site) has a higher percentage of African Americans, a higher percentage of persons and familiesliving below the poverty level, and a lower income (per person) than Orleans Parish andLouisiana.
- A review of the cancer incidence data for Census Tract 1703 from 1983-1997 showed the following:
- A review of breast cancer data showed that:
- No cause or reason for the excess breast cancers in Census Tract 1703 can be determinedfrom this review. The extent of the influence of the established risk factors for breastcancer is not known at this time and no connection with environmental contamination hasbeen made.
From 1988-1997, the observed number of lung, colon, prostate, and total cancerssites combined for Census Tract 1703 are not statistically different from what isexpected as compared with the regional rates.
For the 5-year period, 1983-87, in Census Tract 1703, no statistically significantdifferences between the breast cancers observed and the cancers expected werefound. For all females and black females, the SIRs were less than 1 and for whitefemales the SIR was approximately 2.5.
For the 10-year period, 1988-97, there was a statistically significant 50% excessof breast cancer for all females combined and black females. There was astatistically significant 400% excess of breast cancer observed in white femalesfor Census Tract 1703. The excess in this rate is contributed by the smallpercentage of white females in the census tract and the small numbers that wereused to determine cancer incidence.
For the 15-year period, 1983-97, in Census Tract 1703, no statistically significantexcess of breast cancer was found for all females combined or black females. However, a statistically significant 300% excess of breast cancer did occur inwhite women. Again, the excess in this rate is contributed by the small percentageof white females in the census tract and the small numbers that were used todetermine cancer incidence.
- OPH initially examined the Louisiana Tumor Registry (LTR) data in 1997. Alsoexamined at this time, were birth weights of children born in the area. Follow-upexamination of this information occurred in 2001.
- OPH attended all community meetings to assess community concerns, disseminaterequested information, and answer questions. Explanations of the increase breast cancerrates were explained to the public during these initial meetings. The importance of earlybreast cancer screening was also discussed.
- OPH worked with ATSDR and the Association of Occupation and Environmental Clinicsto provide environmental medical monitoring to participating residents.
- OPH/SEET will continue to monitor the rates of cancer every five years in Census Tract1703.
- OPH/SEET will provide health education to the community on risk factors associatedwith breast cancer if requested. Health education will also be provided on the importanceof early breast cancer screening.
- This health consultation will be placed in the previously established ASL site repository so that residents and stakeholders will have access to the information contained in it.
- Louisiana Office of Public Health, Public Health Assessment for Agriculture StreetLandfill, Public Comment Version, February 1996.
- Beverly H. Wright, The Agriculture Street Landfill Health Survey, Xavier University, 1994.
- U.S. Census Bureau, "1990 Census of Population and Housing", Summary Tape File 1,Technical Documentation.
- U.S. Census Bureau, "1990 Census of Population and Housing", Summary Tape File 1and Summary Tape File 3.
- U.S. Environmental Protection Agency, Remedial Removal Integrated Investigation,Agriculture Street Landfill Site, Dallas, Texas, 1995.
- "Cancer Facts and Figures 1997," American Cancer Society, Inc., 1997.
- Jennifer Kelsey and Marilie Gammon, "The Epidemiology of Breast Cancer," CA-ACancer Journal for Clinicians, Vol. 41, No. 3, May/June, 1991, p.146-165.
- David Schottenfeld and Joseph Fraumeni, Jr. (eds.), Cancer Epidemiology andPrevention, 2nd ed., Oxford University Press, New York, 1996, Chapter 47, "BreastCancer," pp. 1022-1039.
- Mary S. Wolff and A. Weston, "Breast Cancer Risk and Environmental Exposures",Environmental Health Perspectives, Vol.105(Suppl 4), June 1997, p. 891-896.
- US Environmental Protection Agency, Remedial Removal Integrated Investigation,Agriculture Street Landfill Site, New Orleans, Louisiana, 1995, p. 4.1-44.
- Judith S. Mausner and Anita K. Bahn, Epidemiology, An Introductory Text, W.B.Saunders Company, Philadelphia, 1974, p. 310-312.
- Barry L. Johnson (ed.) Advances in Neurobehavioral Toxicology: Applications inEnvironmental and Occupational Health, Lewis Publishers, Michigan, 1990, Chapter 6,"Neurotoxicology in Mexico and Its Relation to the General and Work Environment," p.40.
Louisiana Department of Health and Hospitals/Office of Public Health/Section ofEnvironmental Epidemiology and Toxicology
Erica M. Caesar
Kimberly M. Gallo
Elizabeth G. Kirkland
ATSDR Regional Representative
ATSDR Region VI
ATSDR Technical Project Officer
This Agriculture Street Landfill site, Review of Cancer Incidence Data, health consultation wasprepared by the Louisiana Department of Health and Hospitals under a cooperative agreementwith the Agency for Toxic Substances and Disease Registry (ATSDR). It is in accordance with approved methodology and procedures existing at the time the health consultation was begun.
Tammie McRae, MS
Technical Project Officer, SPS, SSAB, DHAC
The Division of Health Assessment and Consultation, ATSDR has reviewed this public health consultation and concurs with the findings.
Chief, State Program Section, DHAC, ATSDR