Surveying residents

Using chi-squared analyses, proportions of individuals reporting a particular symptom at each site were compared with the combined reports for that symptom at all sites. P indicated how consistently a symptom was reported from site to site, that is, the degree to which numbers of individuals reporting a particular symptom at a given site could be expected based on the frequency of that symptom for all sites combined. Low P values were indicative of symptoms that were largely independent of confounding health-related and environmental factors (e.g., previous medical history, demographics, sources of biosolids, and environmental conditions).

This survey, which did not include an unexposed control group, was primarily intended to describe and document self-reported illnesses. It also provided guidance on which exposure routes (e.g., contaminated drinking water, dermal contact, or inhalation of aerosols) should be the focus of the research. Exposure, on the other hand, was evaluated using a dose-response curve in a separate part of the study concerning environmental assessments.

Assessing environmental conditions

Sludges used at all but one agricultural site (Ontario), according to county records, were limed or applied to alkaline soils then exposed to weathering. Whether sludges used for mining reclamation in Osceola Mills were lime stabilised, however, was disputed. Residents reported that the area was limed after sludging.

County records indicated that biosolids-related complaints for individual patients described in this study were concurrent with land application of Class B biosolids. In the case of one family (Residents 19–22), however, records indicated that dairy wastes, rather than biosolids, had been applied. Unconfirmed reports that dairy wastes had been mixed with Class B biosolids prior to application at this site, however, precluded the authors from ruling out biosolids as a possible cause of the symptoms. According to county officials, soil in this area is naturally alkaline such that irritation similar to that observed with exposure to lime-stabilised sludge may occur.

The NH site was used to develop an approach for assessing exposures to windborne contaminants. At the site, approximately 132 metric tons (dry wt) of Class B biosolids containing 12 metric tons of lime and 318 kg of Stockhausen 644BC cationic polymer (Stockhausen, Inc. Greensboro, NC) were applied to a 10-ha field in late October 1995. The field was located at the end of a paved road with eight houses (A-H) and 28 residents.

Historical records of symptoms experienced in the neighbourhood were kept by one of the residents. Records began approximately 2 weeks after sludges were applied and continued for approximately two years (1995–96). Descriptions of symptoms included coughing, burning eyes, burning throat, congestion, and difficulty in breathing.

To evaluate wind conditions at the site at the time sewage sludge was applied, we used a computer program (WRPLOT Version 3.5, Lakes Environmental Corp., Waterloo, Ontario), topographic information, and historical weather data from Pease Air Force Base located approximately 5 km from the site. Exposure times (h), which were based on duration of wind passing directly from the field toward each house, were determined from wind frequency data averaged for three 8-day periods after sludges were applied (between October 31 – November 23, 1995). Distances (m) each household resided from the nearest edge of the treated field were 23 (A, B); 77 (C), 170 (D); 130 (E); 280 (G); and 320 (H).

Concentrations of air pollutants to which residents were exposed, relative to concentrations over the treated field, were predicted using an Industrial Source Complex (ISC3) Dispersion Model [8]. This prediction was not specific for any pollutant. Instead, it was used as a means for possibly ruling out the effects of ammonia, organic amines, or other gaseous or volatile pollutants. Should airborne chemical contaminants be present only at a fraction of a percent of the concentrations present over the treated field, the result would argue against airborne chemical pollutants as a cause of adverse effects. If, on the other hand, pollutants are sufficiently concentrated to cause burning eyes and other symptoms of irritation at the application site, these same effects may occur among residents in areas where pollutant concentrations are not highly diluted.

The ISC3 air dispersion model is designed to predict concentrations of pollutants off site based on emission rates of the pollutants at their sources and local meteorological conditions. This approach is unreliable when emission rates cannot be accurately determined. Microbiological process rates, which largely control emission rates of airborne pollutants produced by sewage sludges, are highly unpredictable. To circumvent this problem, the model was used only to predict the relative concentrations of air pollutants between the treated field and potentially affected residences. This approach is based on the dispersion of air pollutants as winds carry them over land surfaces, irrespective of their emission rates at the source.

Residents survey

Demographic data for the 48 residents surveyed, along with their symptoms, are listed in Table 1. The average age of the group was 35 years; 56% were female; 23% were smokers. Residents who complained of adverse effects lived within approximately 1 km of the treated fields. Symptoms were primarily associated with sites where freshly applied, lime-stabilised sewage sludges remained on the surface and were exposed to wind erosion.

The primary complaints (a fourth or more reporting the symptom) within 1 h of exposure were coughing (63%, 8 sites), burning throat (56%, 10 sites), burning eyes (56%, 9 sites), and headaches (46%, 6 sites). Within 24 h, residents developed nasal and chest congestion (77%, 7 sites), difficulty in breathing (71%, 8 sites), nausea/vomiting (46%, 7 sites), fatigue (33%, 7 sites), and flu-like symptoms (29%, 5 sites). In some cases, symptoms of chemical irritation (e.g., burning eyes, burning throat) recurred for over a year after applications ceased.

Proportions of individuals reporting coughing and burning throats exhibited the greatest consistency from site to site (P 0.02, 0.03, respectively). Half of the residents reported bacterial, viral or fungal infections within 1 month of exposure (50 %, 8 sites). Most residents also reported symptoms indicating exposure to infectious micro-organisms or microbial products (endotoxins). These included flu-like symptoms (29%, 5 sites), nausea/vomiting (46%, 7 sites), diarrhoea (21%, 4 sites), rashes (25%, 3 sites), and sinusitis (19%, 5 sites).

The most common infections identified by medical records were caused by S. aureus (14 residents at 3 sites, including the additional patients identified in Robesonia). Records indicated that one infection (S. lugdunensis, Resident 44) was probably hospital-acquired. Affected residents in Greenland, NH recalled that respiratory infections experienced by individuals in several households treated by the same physician responded to erythromycin, an antibiotic primarily used for treating gram positive bacterial infections.

In Menifee, CA, two family members complaining of chemical irritation developed different staphylococcal infections. Resident 17 experienced recurrent S. aureus infections of the nasal passages and oral cavity. The spouse was hospitalised for an invasive S. epidermitis infection that developed approximately two days after an undressed surgical incision became contaminated with sand blowing from an adjacent field treated with Class B biosolids.

Mild chronic maxillary sinusitis in one child (Resident 18) living in this household deteriorated into recurrent (unidentified) bacterial infections. The patient responded poorly to oral and IV antibiotic therapy, tonsillectomy, and removal of the adenoids. Rigorous antibiotic therapy, including intramuscular injections of rocephrine, also yielded unsatisfactory results.

County records showed that Class B biosolids were applied to an adjacent field at the time the above-mentioned illnesses developed. The paediatrician treating patients in the area (author C. Pumphrey) noted that children living near the land application sites appeared to develop abnormally persistent sinusitis.

Environmental assessment

Samples of sludges collected at the NH site, which were frozen since 1995, retained a strong faecal odour. Microbiological assays showed that they contained (3.2 ± 3.5) × 10⁵ colony-forming units of bacteria per gram but almost no culturable fungi. Predominant culturable bacteria were identified as common gram-negative environmental isolates and included Alcaligenes faecalis and Brevundimonas diminuta, which occasionally serve as opportunistic human pathogens.

Exposure times (h) for each household when initial symptoms of chemical irritation were reported (Oct 31 – Nov 23) were: 28 ± 25 (A); 48 ± 17 (B, C), 21 ± 12 (D), 88 ± 17 (E); and 13 ± 08 (G, H) (mean ± standard deviation, n = 3). The air-dispersion model indicated that adverse effects from exposure to airborne pollutants from the land application site could not be ruled out as a cause of adverse health effects. Air pollutants at Household C, for example, were frequently present at concentrations of approximately 5 – 40% compared with concentrations 2 m directly above the treated field (Fig. 1).

Numbers of individuals who reported adverse effects compared with the total numbers of individuals per household, respectively, were: 2, 2 (A); 2, 2 (B); 5, 5 (C), 2, 3 (D); 4, 4 (E); 1, 3 (G); and 1, 5 (H). Data from one household (F) were unavailable. Based on a least-squares analysis, proportions of individuals with symptoms increased linearly from 40 to 80 h (r ² 0.98) with time exposed to wind blowing from the field; all occupants in households with exposure ≥ 80 h reported symptoms (Fig. 2). Proportions of individuals with symptoms also decreased linearly with distance from the field from 130 to 320 m (r ² 0.95); all occupants in households living ≤ 130 m from the field reported symptoms.

Prevalence of S. aureusinfections

The second mortality was associated with an outbreak of S. aureus in Robesonia, PA. According to state records, Class B, lime-stabilised sewage sludges from a local wastewater treatment facility were applied to ten fields comprising a total of 300 acres. Applications began in1988 and continued through December 1995. Sewage sludges were applied five times a week at a rate of approximately 1,300 tons (dry wt) per year.

Eight additional individuals lived in four other houses that were also located in close proximity to the field (<50 m). These individuals complained of odours, inundation with houseflies, and respiratory symptoms. They associated these problems with application of biosolids; however, they could not recall having any infections.

Residents reported that feral cats occupying a barn near the treated field had to be destroyed after developing multiple boils and weeping sores after sewage sludge applications began. Similarly, residents at the New Hampshire site reported that one individual and numerous pets (dogs, cats) developed boils after sewage sludges were applied. Pets may have served as a reservoir of staphylococcal infections in these neighbourhoods and, in some cases, transmitted their infections to humans.

Medical records were available for three of the four infected residents: Patients 1, 3, 8. Patient 1, a 17 year-old male with a history of excellent health, was treated for a furuncle of the knee in February1993 then succumbed to S. aureus pneumonia in March 1995 after contracting a rotavirus infection and viral pneumonia. Each of four relatives who frequented the house (≥ 2 d/wk) was also infected with S. aureus. The first of these (Patient 3) developed furuncles concurrently with dermatophytic infections (identified as tinea cruris/tinea corpora). All but two individuals (Patients 6, 7) sought medical attention. These patients were self-treated using antibiotics prescribed for other family members. The last S. aureus infection at Robesonia (Patient 8) occurred approximately two years after biosolids applications ceased.

Overall, the prevalence of S. aureus infections in this study was approximately twenty-five times higher than infections among hospitalised patients, a recognised risk group for S. aureus[9].