Environmental Risk Factors Associated With Pediatric Idiopathic Pulmonary Hemorrhage and Hemosiderosis in a Cleveland Community
Background. Unexplained pulmonary hemorrhage and hemosiderosis are rarely seen in infancy. A geographic cluster of 10 infants with this illness was identified in a large pediatric referral hospital in Cleveland, Ohio, during the period of January 1993 through December 1994. One infant died of severe respiratory failure.
Methods. A case-control study was conducted. Three control infants were matched by age with each case infant. All study infants’ guardians were interviewed. Questions were asked about child care practices and home conditions for the period before case infants’ illnesses. All infants’ records were reviewed, their homes were visited, and a structural and environmental survey was conducted.
Results. All 10 case infants were black, and 9 were male, whereas 50.0% of control infants were male, and 83.3% were black. The case infants’ mean age was 10.2 weeks (range, 6 weeks to 6 months). Matched analysis demonstrated that case infants’ homes were more likely to have had water damage preceding the pulmonary hemorrhage event (odds ratio, 16.25; 95% confidence interval, 2.55 to ∞). Case infants were also more likely to have had close relatives with pulmonary hemorrhage (odds ratio, 33.14; 95% confidence interval, 5.10 to ∞). In addition, 50.0% of case infants experienced recurrent pulmonary hemorrhaging after returning to their homes.
Conclusion. The results of this investigation of a cluster of infants with massive, acute pulmonary hemorrhage and hemosiderosis suggest that the affected infants may have been exposed to contaminants in their homes. Epidemiologic clues, such as water damage in the case infants’ homes, suggest that environmental risk factors may contribute to pulmonary hemorrhage. pulmonary hemorrhage, hemosiderosis, infancy, indoor air pollution, pesticides, volatile organic compounds, cholinesterase.
A cluster of 10 infants with idiopathic pulmonary hemosiderosis (IPH) was identified at a major pediatric referral hospital in Cleveland, Ohio, during the 24-month period from January 1993 through December 1994. One of the infants had died after severe respiratory failure. Pediatric pulmonologists at this hospital had seen only three cases of IPH among children in the preceding 10 years. A case-control study was initiated to identify environmental risk factors for pulmonary hemorrhage among infants.
Spontaneous pulmonary hemorrhage in infants is a rare and dramatic event. In older patients, the pulmonary hemorrhage syndromes, such as Goodpasture’s syndrome,1 are attributed to immune vasculitis and are marked by antiglomerular basement membrane antibodies in serum and in lung and kidney specimens. Other known causes of pulmonary hemorrhage in children and adults include infectious, traumatic, cardiac, and vascular processes.2,3Pulmonary hemosiderosis is the result of chronic, recurrent pulmonary hemorrhage resulting in pulmonary alveolar macrophages becoming hemosiderin laden as phagocytosis occurs in the lung.1,2,4This leads to diffuse deposition of the iron salt hemosiderin within lung parenchymal tissues.
Infants who were previously healthy and then had chronic, recurring episodes of pulmonary hemorrhage of unknown cause are generally categorized as having IPH, the most common form of pulmonary hemosiderosis in infancy and early childhood.2,3 Potential causative factors previously considered for pulmonary hemosiderosis in infancy include immune reactions to cow milk protein5,6 and pesticide exposures.7
Cassimos et al,7 in the first epidemiologic study of IPH published, described 30 cases of IPH that occurred during a 20-year period in northern Greece among children ranging from 1 to 6 years of age. This study made the important observation that the cases occurred in household clusters, where many children were sleeping in rooms near stored grain. The authors suggested that chronic exposure to agricultural pesticides may have been associated with illness. More importantly, however, they also suggested that toxic factors within the home environment may cause illness in families genetically predisposed to that illness. The occurrence in Cleveland of a cluster of infants with IPH provided us the opportunity to examine environmental risk factors that may have contributed to this disease.
Infants included as case subjects were previously healthy and had episodes of acute, diffuse pulmonary hemorrhage of unknown cause during their first 6 months of life during the period from January 1993 through December 1994. All case infants were diagnosed as having pulmonary hemosiderosis based on demonstrating alveolar, hemosiderin-laden macrophages by biopsy or bronchoalveolar lavage 3–6 weeks after the initial hemorrhage. They were all medically treated at the Rainbow Babies and Childrens Hospital, Cleveland, OH. For each case infant, three control infants were selected from the hospital clinic population and from Cleveland’s birth certificate records from the same geographic area as the cluster. Case infants had no known medical problems and were matched to control infants whose dates of birth were within 2 weeks of theirs. Ten case-control groups (30 control infants and 10 case infants) were included in the final study sample.
Medical Record Reviews
All study infants’ medical records were abstracted. Control infants’ birth records were reviewed when available, and case infants’ clinical histories related to the pulmonary hemorrhage events and other admissions were reviewed. Routine perinatal medical data (ie, gestational age, birth weight, and delivery complications) were obtained on each infant, as were results of their newborn metabolic screening tests for sickle cell disease or trait.
All guardians (usually mothers) of study infants were visited at the home by a pediatrician, who administered an in-person questionnaire eliciting information about the infant’s health, infant care practices, and the infant’s home environment. The pediatrician asked the guardians specific questions designed to assess the infant’s exposure to environmental tobacco smoke and illicit drugs. In addition, guardians were asked about their infant’s possible exposure to several classes of toxic agents (eg, pesticides, paints, solvents, and gasoline). Information regarding any home water damage, caused either by plumbing problems, roof leaks, or flooding, was obtained for the 6-month period preceding the case infants’ admissions to the hospital for pulmonary hemorrhage. The interviews were conducted in the homes where the case infants lived at the time they became ill.
During the initial home visit, while the interview was being conducted, a preliminary environmental survey was performed by a local registered sanitarian. The sanitarian evaluated the general condition of the home and recorded any visible signs of water damage to the physical structure. A complete household inventory of toxic chemicals was obtained.
An environmental sampling strategy, developed by the National Institute for Occupational Safety and Health (NIOSH), was used to evaluate indoor air and sample environmental surfaces for the presence of organophosphate, carbamate, and pyrethrin classes of insecticides. Four teams consisting of registered sanitarians and environmental hygienists were trained by an industrial hygienist from the NIOSH to conduct the pesticide-sampling protocol. Area air sampling for pesticides (organophosphates, carbamates, and pyrethrins) and volatile organic compounds (VOCs; 8 to 10 carbon fluorocarbons of common household airborne solvents) was conducted in each home in either the room where the infant slept or the area most commonly occupied by the mother and infant. Wipe samples were collected in areas thought likely to have received insecticidal treatment. These areas included baseboards in the kitchen; the floor area below the kitchen sink; the base of the tub, toilet, and vanity in the bathroom nearest the infant’s bedroom; and the baseboard in the room in which the infant usually slept.
Organophosphate pesticides were sampled by using SKC Occupational Safety and Health Administration Versatile Sampling tubes (SKC Inc, Eighty Four, PA) connected to Gillian personal sampling pumps.
Surface samples from 10 × 10-cm areas were collected with 4 × 4-in Johnson & Johnson sterile gauze pads wetted with commercially available 70% isopropyl alcohol. Investigators wore disposable polyethylene gloves previously determined not to contain pesticide residues. They changed gloves between samples to avoid cross-contaminating the samples or the field blanks.
Environmental pesticide analyses were conducted at the NIOSH in Cincinnati, OH. Because of method specificity, the same method could not be used to analyze organophosphates, carbamates, and pyrethrins. The wide variety of pesticides available necessitated that air and wipe samples be separated for the various analyses. An analytical screen was performed to detect the presence of organophosphates, carbamates, and pyrethrin pesticides. Air samples collected during the investigation were analyzed for organophosphate pesticides by NIOSH Method 56008 with a Hewlett Packard HP5890 gas chromatograph equipped with a flame photometric detector.
Biological Sample Collection and Analysis
All infants in the study had blood specimens drawn and analyzed at the University Medical Laboratories of Cleveland. All samples were obtained from case infants during the initial or recurrent episodes of pulmonary hemorrhage (acute phase of the illness). Using standard laboratory procedures, analysts tested for complete blood cell count and reticulocyte count and conducted a glucose-6-phosphate dehydrogenase screen and acetylcholinesterase (ACHE) and cholinesterase (CHE) assays. A blood specimen from each infant was sent to the IBT Reference Laboratory (Lenexa, KS) where milk allergen-specific immunoglobulin G was measured by enzyme immunoassay in a microtiter plate format.
All infants had urine specimens collected, frozen, and sent for analysis to the Centers for Disease Control and Prevention, National Center for Environmental Health, Division of Environmental Health Laboratories Sciences (Table 1). There, analysts performed a quantitative urinary creatinine determination on a Kodak Ektachem 250 Dry Chemistry Analyzer, using a single-slide, two-point enzymatic creatinine test (Eastman Kodak Co, Rochester, NY).9 The intended purpose of creatinine determinations was for the correction of other urinary analytes (ie, pesticides).
A commercially available radioimmunoassay technique (Roche Abuscreen kit 43243; Roche Diagnostics, Nutley, NJ) was used to analyze urine specimens for the presence of the cocaine metabolite benzoylecgonine. The nicotine metabolite cotinine was determined with a microplate cotinine enzyme immunoassay kit from STC Diagnostics (Bethlehem, PA).
Urinalysis for organophosphate pesticides was performed by using capillary gas chromatography combined with tandem mass spectrometry. The analysis was for 12 selected pesticide residues representing potential exposure to more than 30 pesticides.10 The method involved the use of a 1-mL urine sample, which underwent enzymatic hydrolysis, pH adjustment, and extraction. The extract was then concentrated and back-extracted with base before derivatization, clean-up, and concentration for analysis with gas chromatography combined with tandem mass spectrometry by the isotope dilution technique (carbon 13-labeled internal standards).
The questionnaire and biological sampling were all performed under a protocol approved by the University Hospitals of Cleveland Institutional Review Board.
Data were entered, and a preliminary unmatched analysis was performed with Epi Info version 6.0 (Centers for Disease Control and Prevention, Atlanta, GA). Later epidemiologic, environmental, and biological data were merged by using SAS Systems (SAS Institute Inc, Cary, NC), and a univariate, matched analysis was performed. Odds ratios (ORs) and 95% exact confidence intervals (CIs) were calculated. Last, multivariate analysis was performed by using conditional logistic regression models constructed with the LogXact System (Cytel Software Corp). This method enabled us to determine the effect of multiple potentially confounding variables. Significance was defined as aP < .05 and CIs that excluded unity.
For most case infants, acute pulmonary hemorrhage episodes began with a brief prodrome heralded by an abrupt cessation in crying, limpness, pallor, and/or color change (Table 2). This prodrome was followed by acute hemoptysis, lethargy, grunting, and respiratory failure in most infants. All infants required admission to the pediatric intensive care unit, where they received mechanical ventilation for an average of 5 days. Most infants demonstrated diffuse, bilateral, alveolar pulmonary infiltrates by chest roentgenogram.
Table 3 summarizes the descriptive epidemiology of the study infants and their mothers. The case infants’ mean age was 10.2 weeks (range, 6 weeks to 6 months). All control infants were aged matched within 2 weeks to the case infants’ dates of birth. The makeup of the case group differed from that of the control group in that the case group consisted of 90.0% male and 100.0% black infants, whereas the control group had only 50.0% male and 83.3% black infants. At birth, case infants had a mean gestational age of 36.6 weeks (range, 27.0 to 41.0 weeks), and control infants had a mean gestational age of 39.2 weeks (range, 32.0 to 41.0 weeks). Similarly, case infants’ mean birth weight (2.6 kg) was also lower than that of control infants (3.4 kg). Case and control infants’ mothers were similar with respect to their mean ages (21.2 and 24.3 years, respectively), the percentage of them receiving Medicaid (80.0% and 83%, respectively), and their mean level of education years completed (11.4 and 11.5 years, respectively).
Results of laboratory analysis of biological specimens from case infants (Table 4) were consistent with an acute hemorrhage. The complete blood cell count demonstrated a mean hematocrit of 24.0% for case infants and 32.6% for control infants (P < .05), a mean hemoglobin (Hgb) concentration of 8.3 g/dL for case infants and 11.2 g/dL for control infants (P < .05), and a total red blood cell count of 3.5 million/mm3 for case infants and 4.2 million/mm3 for control infants (P < .05). Blood smears from case infants examined under bright-field microscopy were estimated to have 1+ to 2+ hemolysis, suggesting that a microangiopathic hemolytic process was involved. The mean corpuscular volume and mean corpuscular Hgb were within normal limits, excluding a microcytic or hypochromic type of anemia such as chronic iron deficiency.
Mean white blood cell counts were 12 300/mm3 for case infants and 10 300/mm3 for control infants. Both counts were within normal limits for the infants’ ages. Results of coagulation studies were all within normal limits for case infants, including a mean platelet count of 351 000/mm3, a prothrombin time of 13.5 seconds, and a partial thromboplastin time of 40.4 seconds. The mean platelet count for control infants was 412 600/mm3. Prothrombin time and partial thromboplastin time were not measured for control infants.
Mean CHE and mean ACHE levels were within normal limits for both case and control infants (6.8 vs 8.5 kU/L for CHE, and 26.5 vs 28.6 U/g of Hgb for ACHE, respectively).
Control infants had significantly higher levels of milk protein allergen-specific immunoglobulin G than did case infants (54.7 vs 15.1 units/mL; P < .001).
Results of quantitative glucose-6-phosphate dehydrogenase screening were within normal limits for all case and control infants. Newborn metabolic screening results for sickle cell disease were also normal for all case and control infants, except for one control infant whose results were positive for sickle cell trait.
During laboratory analysis of urine samples, only a subset of 15 specimens (5 from case infants and 10 from control infants) was assayed for organophosphate analytes, VOCs, cotinine, and cocaine (benzoylecgonine) metabolites. Results for pesticides and VOCs were either below the limits of detection (LOD) or not significantly above the LOD. Twelve (80.0%) of 15 case and control infants tested had urinary cotinine levels greater than the 2.0 ng/mL LOD. One case and 2 control infants had levels greater than 40 ng/mL.
The 10 case infants had a total of 21 episodes of acute pulmonary hemorrhage during a 24-month period from January 1993 through December 1994 (Fig 2). Of these episodes, 11 were recurrent pulmonary hemorrhage episodes in 5 case infants. All infants with recurrent episodes required readmission to the hospital, and 1 infant died after multiple recurrences. Recurrent hemorrhaging on several occasions occurred within 48 hours after the infants were discharged into the homes where they became ill. As depicted in the epidemic curve, the cluster seems to have started in the late fall of 1993, with no cases having occurred in the summer of 1993 and only 1 case having occurred before this in that same year. In addition, 50.0% of the pulmonary hemorrhage episodes occurred during the months of June through October (2 episodes in October 1993 and 8 episodes from June through September 1994). Of the remaining episodes, 38.0% occurred in the winter months (4 episodes from November 1993 through January 1994 and 4 episodes from November through December 1994).
All case infants lived within a geographic area defined by six ZIP codes (Fig 1), which primarily includes East Cleveland and the northeast segment of the city of Cleveland, two areas having older and less maintained housing and overflow cross-connected sewer systems. Two natural brooks, Doan Brook and Dugway Brook, flow through the cluster area, and the larger one (Dugway Brook) divides the cluster area almost in half. These natural brooks are prone to flooding from combined sewer overflows during heavy rainfalls. During the summer of 1994, the National Climatic Data Center reported some of the heaviest rainfall in the history of Cleveland.11 For August 13, 1994, the Northeastern Ohio Regional Sewer District rain accumulation report indicated that 4.21 in of precipitation were measured in a rain gauge present in the cluster area.12 The regional sewer report recorded on that day a daily storage capacity versus overflow ratio of 4.39 million to 15.78 million gallons, a significant excess of almost 12 million gallons that resulted in serious waterway overflows into surrounding, low-lying streets and homes in the cluster area.
After we conducted a matched analysis with conditional logistic regression, three variables remained significant (Table5). Case infants were more likely to be male (OR, 9.12; 95% CI, 1.05 to ∞). In addition, case infants were more likely to have had a close relative who also coughed blood while living in the same home (OR, 33.14, 95% CI, 5.10 to ∞). The greater birth weight of control infants was of borderline significance and may suggest a protective effect against pulmonary hemorrhage (OR, 0.12; 95% CI, 0.01 to 0.65).
Case infants were also significantly more likely to have had any water damage to their homes from flooding after heavy rainfalls, from roof leakage, or from water pipe leakage saturating plaster walls and ceiling tiles during the 6-month period preceding the infants’ pulmonary hemorrhage episodes (OR, 16.25; 95% CI, 2.55 to ∞). In all case infants’ homes, water damage had occurred within 1 month before the infants illness and in many cases was present during their first episode of pulmonary hemorrhage. Furthermore, there was no attempt in any case infant’s home to clean up water damage, and often the source of water contamination was never repaired. However, clean-up efforts were reported in all 7 of the control infants’ homes in which water damage was reported. No significant differences were found between the types of heating and cooling systems used in case and control homes. However, unmatched univariate analysis demonstrated that in 7 (87.5%) of 8 case infants’ homes but in only 11 (45.8%) of 24 control infants’ homes electric fans were reported to have been used regularly to ventilate infant living environments (OR, 8.27; 95% CI, 0.77 to 208.51).
Two risk factors did not attain statistical significance after stratification in a model that included water damage: any smoking in the home and having ever breastfed the infant. Although case infants were almost eight times (OR, 7.9; 95% CI, 0.9 to 70.6) as likely to have had any smoking in the home before their illnesses, the 95% CI included unity. Smoking was reported in 7 (70.0%) of 10 case infants’ homes that had water damage but in only 2 (28.6%) of 7 control infants’ homes that had water damage.
Having ever been breastfed seemed to be protective (OR, 0.16; 95% CI, 0 to 1.16); however, again significance was not attained. No case infant was breastfed. Eleven (36.6%) of 30 control infants were breastfed, of which 4 of 7 (57.1%) reported both water damage and breastfeeding, and 3 of 7 (27.2%) reported smoking, breastfeeding, and water damage.
Environmental air and wipe sampling revealed trace amounts of commonly used household pesticides (chlorpyrifos, diazinon, and other carbamate or pyrethrin compounds) in several case and control infant homes. The one deceased infant’s home contained the largest concentration of airborne solvents and pesticide residues, specifically chlorpyrifos preparations. This sample was reported by the laboratory at the limit of quantitation for chlorpyrifos, or 0.14 μg/sample. The LOD for chlorpyrifos was reported as 0.04 μg/sample. The reported concentration represents a semiquantitative or trace value, because it lies between the range from the LOD and limit of quantitation values. A minimum detectable concentration for chlorpyrifos of 0.2 μg/m3 was calculated on the basis of a 228-L sample volume.13 The remainder of the samples were reported below the LOD for all other pesticide analytes tested.
The observed greater frequency with which case infants’ homes experienced water damage (0R, 16.25; 95% CI, 2.55 to ∞) and the observation that case infants’ onset of illness occurred soon after home water damage both support a hypothesis in which the infants’ home environments were an important factor in the causal chain leading to IPH.
The epidemiologic evidence presented in this investigation supports the notion that indoor, environmental risk factors may exist in the homes of infants with IPH. Included in this evidence was the fact that some recurrences of pulmonary hemorrhage occurred after case infants returned to their homes. In addition, one case infant’s twin sibling and one case infant’s mother had pulmonary hemorrhage while living in the same homes where the case infants became ill (OR, 33.14; 95% CI, 5.10 to ∞). The one twin sibling of the case infant underwent elective screening bronchoscopy and was found to have hemosiderin-laden macrophages, the marker for chronic pulmonary hemorrhage. The other case infant’s mother, when she was a 6-week-old infant, had massive hemoptysis of unexplained origin while living in the same home.
Although Sherman et al16 had described four infants with pulmonary hemorrhage in Cleveland in a previously published case report, this is the first report of a temporal and geographic cluster of IPH occurring among infants. Such a cluster of a rare and idiopathic pediatric illness provides a unique opportunity to identify risk factors.
The clinical presentation of the infants with acute pulmonary hemorrhage in this cluster was not consistent with acute organophosphate poisonings seen in young children. Accidental poisonings with organophosphate pesticides among pediatric populations are well described in the scientific literature.17,18 Fatal organophosphate pesticide exposure is known to occur through dermal, oral, and inhalational routes. High-dose exposure to parathion, an organophosphate pesticide approved for agricultural use only, has been demonstrated to cause acute pulmonary edema among pediatric patients.17 Although acute pulmonary hemorrhage is not commonly seen with pesticide poisoning, pulmonary hemorrhage and hemolytic anemia have been described among adults with toxic exposure to trimellitic anhydride, an industrial compound used for curing epoxy resins.20
Our investigation is the first-known comprehensive evaluation of the indoor air and surface environment for organophosphate pesticides and VOCs in the homes of infants with IPH. We demonstrated no significant difference between case and control infants with respect to their exposure to indoor air and surface levels of pesticides and VOCs. Although the samples from case infants were generally obtained long after the onset of illness, several of the case infants’ homes were tested within days after the onset of illness. Even in these homes, levels of pesticides and VOCs were less than the LOD.
The adverse perinatal effects of maternal tobacco and cocaine use have been well documented.21 Increased morbidity and mortality from perinatal cocaine exposure are associated with lesions of vascular disruption, including placental abruption and perinatal cerebral infarction.21,24 Pulmonary hemorrhage in adults has been previously attributed to inhalational exposure to tobacco.28 Case infants in our study were eight times more likely than control infants to have been exposed to environmental tobacco smoke, and their elevated risk approached statistical significance. It is plausible that infants exposed to chronic low levels of environmental tobacco smoke could be at greater risk for respiratory cell death secondary to the cumulative effect of exposure to combined respiratory toxins. Many study infants’ guardians reported that multiple persons smoked regularly in the infants’ indoor home environments. The urinary cotinine measurements were performed, in most instances, substantially after case infants’ pulmonary hemorrhage events; we did not have an assessment of the level of these toxicants in these infants at the time of their illnesses.
Early studies by Heiner et al6 attributed IPH of infancy to circulating antibodies against cow milk proteins, which they found in the blood of children with IPH. However, this hypothesis has not been supported by either a demonstration of immune complexes in the lung parenchyma by immunofluorescent techniques or by delineation of a pathophysiologic mechanism. In addition, the hypothesis has not been previously tested in a case-control study design, and it was not supported by the data from this investigation. In fact, we found that control infants had significantly higher levels of milk protein-specific immunoglobulin G than did these case infants (P < .001). However, we recognize that pulmonary hemosiderosis associated with cow milk allergy is a different entity than the cases described in this investigation.
This environment-based model describes an interaction between a host (case infant), who may be biologically or genetically predisposed to illness, and the home environment, which may contain multiple toxins to which the case infants may be exposed. There were no differences between case and control infants with respect to having a genetic blood disorder (ie, sickle cell disease, sickle cell trait, or glucose-6-phosphate dehydrogenase deficiency); however, other possibly unidentified genetic predispositions may have existed.
Infant lung development during the first 6 months of life is characterized by rapidly multiplying, immature alveoli, pulmonary arteriole proliferation, and continued differentiation of alveolar lining into type I and II epithelial cells.29 These factors could increase an infant’s susceptibility to inhaled environmental toxins. The male preponderance identified in this cluster, however, is difficult to explain, although male infants are reported to have greater lung instability and increased risk for neonatal respiratory distress syndrome.29,30
Another precipitating factor that might alter an infant’s likelihood of disease is the lack of breastfeeding, however, this variable was limited by the small sample size included in our cluster. Therefore, the study power to detect a difference between case and control infants for these variables was limited. Nonetheless, breastfeeding could have conferred a protective advantage against pulmonary hemorrhage among infants. Breast milk provides immunoglobulin A, antiinflammatory, and immunomodulatory factors, which in turn provide breastfed infants enhanced immunity and a demonstrated reduction in gastrointestinal and respiratory illnesses in the newborn period.31,32 This same enhanced immunity could improve an infant’s response after exposure to environmental toxins.
The cluster of IPH cases that we examined, although not directly attributable to any one specific environmental factor, strongly points toward an agent-host-environment interaction that is influenced by several risk factors. The strength of the association between home water damage and IPH suggests that the primary environmental risk factor for IPH may be associated with water damage.
We are grateful to Eric J. Esswien MSPH, CIAQP, industrial hygienist from the NIOSH, Industrial Hygiene Section, Hazard Evaluations and Technical Assistance Branch, Division of Surveillance, Hazard Evaluations and Field Studies, for his expert professional assistance in training for and performing environmental testing. We are grateful to David R. Olson, PhD, Robert Hill, PhD, Harry Hannon, PhD, and Eric Sampson, PhD, from the National Center for Environmental Health, Centers for Disease Control and Prevention, for providing statistical and laboratory assistance. We are also grateful to Michael D. Infeld, MD, Paul Smith, DO, and Connie Judge, MD, of the Rainbow Babies and Childrens Hospital for providing their professional assistance during the study. We are also indebted to the dedicated staffs of the Cuyahoga County Board of Health, Environmental Health and Toxicology Branch, the City of Cleveland Department of Public Health, and B. Kim Mortensen, PhD, Kim A. Winpisinger-Slay, MS, Steven A. Wagner, MPH, and S. Amanda Burkett, MA, from the Division of Epidemiology, Ohio Department of Health, who provided scientific, technical, and administrative support during the investigation.
- Received December 28, 1995.
- Accepted April 23, 1996.
Reprint requests to (R.A.E.) Air Pollution and Respiratory Health Branch, National Center for Environmental Health, Centers for Disease Control and Prevention, 4770 Buford Hwy NE, Mail Stop F-39, Atlanta, GA 30341–3724.
- IPH =
- idiopathic pulmonary hemosiderosis •
- NIOSH =
- National Institute for Occupational Safety and Health •
- VOC =
- volatile organic compounds •
- ACHE =
- acetylcholinesterase •
- CHE =
- cholinesterase •
- OR =
- odds ratio •
- CI =
- confidence interval •
- Hgb =
- hemoglobin •
- LOD =
- limit of detection
- ↵Levy J, Wilmott R. Pulmonary hemosiderosis. In: Hilman BC, ed.Pediatric Respiratory Disease: Diagnosis and Treatment.Philadelphia, PA: WB Saunders Co; 1993:543–549
- Heiner DC,
- Sears JW,
- Kniker WT
- ↵National Institute for Occupational Saftey and Health (NIOSH). Organophosphorus pesticides, method 5600. In: Eller PM, ed. NIOSH Manual of Analytical Methods. 4th ed. Cincinnati, OH: NIOSH; 1994:94–113
- ↵Eastman Kodak Inc. Kodak Ektachem 250 Manual of Operations. Dry Chemistry Analyzer. Rochester, NY: Eastman Kodak Inc; 1992. Publication MP2-49
- Hill RH,
- Shealy DB,
- Head SL,
- et al.
- ↵National Oceanic and Atmospheric Administration, United States Department of Commerce. National Climatic Data Center Report. Cleveland, OH: National Weather Service Forecast Office; 1993–1994
- ↵Northeastern Ohio Regional Sewer District, Sewer Control Systems.Daily Flow Report. Cleveland, OH: Northeastern Ohio Regional Sewer District; 1993–1994
- ↵Vacarro JR. Risks associated with exposure to chlorpyrifos and chlorpyrifos formulation components. In: Racke KD, Leslie AR eds.Pesticides in the Urban Environment: Fate and Significance.Washington DC: American Chemical Society; 1993:303
- American Conference of Governmental Industrial Hygienists (ACGIH): Documentation of threshold limit values (TLVs) and biological exposure indices. In: Chlorpyrifos. 6th ed. Cincinnati, OH: American Conference of Governmental Industrial Hygienists Inc; 1991;1:310–311
- Bledsoe FH,
- Seymour EQ
- Shannon MW,
- Hite C,
- Woolf A
- DeKun L,
- Daling JR
- Kleinman JC,
- Mitchell B,
- Pierre JR,
- Madans JH,
- Land GH,
- Schramm WF
- Malloy MH,
- Kleinman JC,
- Land GH,
- Schramm WF
- Lowry R,
- Buick B,
- Riley M
- ↵Harding R, Bocking AD. Fetal lung growth. In: Meisami E, Timiras P, eds. Handbook of Human Growth and Developmental Biology.Boca Raton, FL: CRC Press Inc; 1990;3:131–148
- ↵Kuhn C, Askin FB. Lung and mediastinum. In: Kissane JM, Anderson WAD, eds. Anderson’s Pathology. 8th ed. St Louis, MO: CV Mosby Co; 1985;1:833–897
- Howie PW,
- Forsyth JS,
- Ogston SA,
- Clark A,
- Du V,
- Florey C
- ↵Taylor JK. In: Quality Assurance of Chemical Measurements.Chelsea, MI: Lewis Publishers; 1987:78–84
- Long GL,
- Winefordner JD
- Copyright © 1997 American Academy of Pediatrics