NPDS is a national database and data management system owned by the American Association of Poison Control Centers (AAPCC). It receives information on calls made by the public to the 57 PCs across the US in near, real-time. We used a cross-sectional study to identify all pediatric lamp oil product exposures reported to PCs from 1/1/2000 to 12/31/2010 by reviewing NPDS data. PC and NPDS data use a system of codes and standardized outcome categories (Table 1) to track potentially hazardous exposures. We defined a lamp oil product exposure as any reported exposure with the AAPCC generic code for lamp oil exposures (201031). Aggregate data on daily exposure call count, age, gender, state where exposure occurred, date of exposure, site of exposure, medical outcome, management site, clinical effects, and therapies were collected and analyzed using Microsoft (MS) Excel 2010 and SAS 9.3. PC records/notes were not available for review. PC penetrance describes the rate of exposures called to PCs of the implicated substance per year per unit population. This measure was used to compare regions of the country and determine which areas have higher reported exposures given a standard population. For the regional penetrance analyses, states were first grouped into four major regions according to the US census (Fig. 1): Northeast (Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, Connecticut, New York, Pennsylvania, and New Jersey); Midwest (Wisconsin, Michigan, Illinois, Indiana, Ohio, Missouri, North Dakota, South Dakota, Nebraska, Kansas, Minnesota, and Iowa), South (Delaware, Maryland, District of Columbia, Virginia, West Virginia, North Carolina, South Carolina, Georgia, Florida, Kentucky, Tennessee, Mississippi, Alabama, Oklahoma, Texas, Arkansas, and Louisiana) and West (Idaho, Montana, Wyoming, Nevada, Utah, Colorado, Arizona, New Mexico, Alaska, Washington, Oregon, California, and Hawaii). We then compared lamp oil product exposure call volume among regions. Next, region-specific PC penetrance was determined by dividing the total number of region-specific lamp oil product exposures by the corresponding region’s total estimated pediatric population (<5 years of age) from 2000 to 2010, multiplied by 10,000.¹² A chi-square analysis was conducted to investigate the association between penetrance and region. Seasons were defined as: Winter (December, January, and February), Spring (March, April, and May), Summer (June, July, and August), and Fall (September, October, and November). The number of lamp oil product exposure calls occurring in each season from 2000 to 2010 was then determined. Poisson regression was used to model the count of exposures per day by season. The number of exposures occurring the day before, on, and after Independence Day (July 4th) from 2000 to 2010 was compared to the count during the remaining days in July also using Poisson regression. All analyses were conducted using SAS v9.3.

The PCs of Utah, Wisconsin, New Jersey, Kentucky, and Cincinnati, Ohio each obtained approval from their Institutional Review Board (IRB) to participate in telephone survey component of the study. A lamp oil product was defined as a product that is used as fuel for lamp oil-based lamps, which was verified by reviewing the call narrative. The PC-based co-investigators used the following codes to identify potential lamp oil product ingestions: lamp oil (0201031), other types of hydrocarbons (0039510), and miscellaneous essential oils (0077360). PC staff then selected the 10 most recent calls from the study period of 1/1/2007 to 12/31/2008. Each PC enrolled callers beginning with the latest call, until 10 eligible calls had been consented for participation. The same questionnaire was administered by all the PCs and is available upon request. Inclusion criteria for study enrollment included: (1) age ≤5 years; (2) an exposure to lamp oil or associated product (e.g, tiki torch fuel); and (3) an exposure route of ingestion (including aspiration). De-identified results were provided to CDC co-investigators for analysis. Data were analyzed using SAS v9.3.

From 2000 to 2010, 23,536 pediatric exposure calls about lamp oil products were reported to NPDS. A mean of six exposure calls per day was reported across all PCs (range: 1–19). The mean age of exposed children was 2 years, median 20 months, and range 13 days to 5 years. Most reported exposures involved males (n = 13,758; 59%). Table 2 illustrates the frequency distributions for exposure site, management site, medical outcome, clinical effects, (e.g., signs, symptoms, laboratory abnormalities, etc.) and therapies used according to standard NPDS classifications.¹³ The most common site of exposure was the patient’s home (n = 21,510; 91%) and most did not require medical treatment in a health-care facility (n = 14,506; 62%). Most had either no or minimal clinical symptoms (nausea, vomiting, etc.) as a result of the exposure (n = 19,739; 84%). The proportion of pediatric lamp oil product exposure calls reporting death as the outcome in NPDS data for the study period was 0.03% (7 deaths) and the proportion that developed symptoms was 45%. The most common reported adverse health effect was cough/choke (n = 8,027; 34%). The most common therapy given was dilute/irrigate/wash (n = 15,179; 64.5%). The number of lamp oil product exposures reported to NPDS differed significantly by region (p value < 0.0001). The Midwest had observed exposure frequencies that were 20.5% higher than expected on the basis of the null hypothesis of no association between region and exposures and the Northeast had observed frequencies that were 22.6% lower than expected. Observed frequencies in the South and West were 0.58% and 2.2% lower than expected, respectively. The Midwest had the highest penetrance of all the regions (1.08 exposures per 10,000 population). The Northeast had the lowest of the regions (0.69 exposures per 10,000).

The mean count of reported lamp oil product exposures per day by season from 2000 to 2010 was as follows: Fall, 6.0; Summer, 6.4; Spring, 5.7; and Winter, 5.6. When comparing the daily seasonal mean values, there was a statistically significant difference (p < 0.001) for every comparison except between Spring and Winter (p = 0.30). The number of reported lamp oil product exposures around July 4th was significantly higher (2.09 vs 1.89 calls/day, p < 0.002) compared to the number of exposures for the rest of the month of July.

Ten telephone surveys were completed from each PC for a total of 50 surveys. However, 11 surveys for ingestional exposures to air fresheners, herbal medications, perfumes, stove fuels, lubricants, and furniture grease were excluded from the analysis. Demographic information is summarized in Table 3. The median age was 2 years. Most patients were male (n = 22; 56%). Exposure information and circumstances surrounding ingestion can be found in Table 4. When asked about place of ingestion, more than half of the respondents reported the ingestions occurred in the child’s house (n = 22; 56%). The remainder of ingestions took place on a deck/patio/porch (23%), backyard (15%), public place (2%), or garage (2%). The majority of the children were not alone at the time of ingestion (n = 30; 77%) and reportedly drank less than a mouthful (n = 24; 62%). Most ingestions occurred in the afternoon (defined as between 12:00 and 17:59; n = 15, 38%) and evening (defined as between 18:00 and 23:59; n = 15, 38%).

The majority of callers reported that the ingested product was not stored close (within 1–3 feet) to food (n = 35; 90%). None of the 39 children could read at the time the ingestion occurred and only one was able to recognize warning symbols. Most of the children drank the lamp oil product from the original container (n = 23; 59%) and most of the original containers did not have a child-resistant closure (n = 14; 61%). For the 16 in which the exposure was associated with a non-original container, the lamp oil or related product was stored in either a lamp, candle, or tiki torch.

Among all lamp oil products, tiki torch fuel was the most commonly ingested product type (n = 13; 33%). A table or counter was the most commonly reported storage location (n = 6; 15%). When asked about lamp oil color, 18 (46%) reported a presence of a color and out of these, yellow or gold was the most common color reported (n = 12; 67%). Only 9 (23%) described the products’ smell as pleasant. The majority of children (n = 24; 62%) ingested less than a mouthful, and 3 (8%) ingested more than a mouthful. Of these three, none described the product’s smell as pleasant. The amount of lamp oil ingested did not appear to be associated with either the smell of the oil (p = 0.19) or the presence of color (p = 1.00) in this small sample. One child had a previous history of pica and approximately one-third of the children had a previous history of eating non-food items (n = 15; 38%). Eleven (28%) callers reported previous contact with their PC or their doctor because their child had eaten a non-food item on a previous occasion. None of the children who had previously reported non-food item ingestions were able to recognize warning or hazard symbols. Approximately half (n = 5; 45%) of these children had a non-food eating habit as reported by the respondent. Approximately half of the children who ingested a lamp oil product had no symptoms after ingestion (n = 18; 46%). The maximum number of symptoms reported was six. All of the children who reported symptoms after exposure developed at least one symptom within 4 h. In two of these instances, however, the initial symptoms reported were choking and/or vomiting with more serious symptoms (cough and fever) occurring 5–12 h after the ingestion. The local PC was the agency/healthcare provider most often notified first after an ingestion (n = 33; 85%). Of those that initially called an agency/healthcare provider (n = 36), approximately half reported no symptoms (n = 17; 44%). Among all children, 12 (31%) sought care at a healthcare facility (HCF) defined as a hospital emergency department (ED), pediatrician’s office, or urgent care center. Two children sought care though they were asymptomatic. Most of the children who did not seek care drank less than a mouthful (n = 19; 49%). Of those who sought care, half drank a mouthful or more (n = 6; 50%) and most eventually went to the local ED (n = 9; 75%).

Of the 12 children who sought care at a HCF, ten (83%) received treatment or diagnostic studies and therapies described in Table 5. The average length of stay in the ED reported by survey respondents was 3.1 h (range: 0–8 h; mode: 2 h). Of these ten, three (30%) were eventually admitted. The mean length of admission reported was 1.7 days (range: 1–2 days). Thirty three of those surveyed provided answers regarding current health status of the affected child. Two (6%) had a history of asthma. Five (15%) had a previous history of breathing problems. One (3%) reported chronic breathing problems after lamp oil exposure and one (3%) reported development of a persistent cough that had resolved by the time of this survey. Three (10%) children were placed on medication(s) by a physician as a result of lamp oil or related product-related illness after the acute event. These medications included albuterol inhaler, steroid inhaler, dexamethasone, and antibiotics. It is unknown whether these medications were given in the inpatient or outpatient settings. At the time of the survey, none of the children required continued treatment with any of these medications.