Geographic Distribution of Scorpion Exposures in the United States, 2010–2015

Abstract

Objectives. To determine the geographic distribution of scorpion envenomations in the United States by zip code, with particular attention to the neurotoxic Centruroides sculpturatus (Arizona bark scorpion), for which an antivenom is available.

Methods. We obtained scorpion exposure cases for 2010 to 2015 from the National Poison Data System. Using geographic information systems software, we mapped total exposures and incidence rates for 9 states that reported more than 100 annual calls. We also mapped cases that reported fasciculations and nystagmus (unique to C. sculpturatus among native scorpions).

Results. The highest exposure incidences occurred in Phoenix (up to 677 per 100 000 population) and Tucson (584), both in Arizona. Elsewhere, high incidences were found in El Paso, Texas (213); Oklahoma City (209) and Tulsa (178), Oklahoma; and Las Vegas, Nevada (170). Fasciculations and nystagmus were reported in Arizona and southeastern Nevada, with small numbers in surrounding states, including Utah.

Conclusions. Scorpion exposures occurred at baseline rates throughout many of the southern states, whereas several states reported effects indicative of Arizona bark scorpion envenomation.

Public Health Implications. Public and health care provider education, as well as the stocking of antivenom, should be targeted based on these findings.

Over the past decade, nearly 17 000 scorpion exposures were reported annually to poison control centers (PCCs) throughout the United States.¹ Although Arizona accounted for a large proportion of these, with 11 500 exposures per year, Texas, Nevada, and Florida all averaged more than 500 calls per year. In addition, 5 other states reported at least 100 calls per year. These are likely underestimates because scorpion exposures do not require mandatory reporting. Overall, 16 275 (8.8% of all US scorpion exposures) patients were referred to or presented at a health care facility between 2005 and 2015. Although there are many native scorpion species, clinical effects of envenomation by most is generally limited to pain, paresthesia, and tissue swelling at the sting site. The one native species of medical concern in the United States is Centruroides sculpturatus (Arizona bark scorpion) because its neurotoxic venom not only causes pain (without local skin reaction), but involuntary eye and limb movements, as well as trouble swallowing and controlling secretions.² This inability to control oral secretions is the most concerning effect, particularly in children, because the resulting respiratory distress can be severe enough to necessitate endotracheal intubation and mechanical ventilation. Although an antivenom is available to treat severe toxicity,³ its use has been affected because of cost concerns.^4,5

Because most (97.8%) scorpion exposures reportedly occur in or around the home,¹ it may be possible to reduce the number of envenomations through public education. Because of its greater potential toxicity, such education would be especially important in areas where C. sculpturatus is found, historically believed to be Arizona and small parts of California, Nevada, New Mexico, and Texas.² To help guide prevention efforts, we undertook this study with 2 objectives: (1) to characterize the geographic distribution of scorpion stings in the 9 US states with more than 100 annual reported scorpion exposures, and (2) to map the current distribution of exposures that are most consistent with C. sculpturatus based on clinical effects. Because the overall intention was public education, we used geographic information systems (GIS) software to generate these maps to better visualize the areas involved.

METHODS

The National Poison Data System (NPDS; http://www.aapcc.org/data-system) is a database owned by the American Association of Poison Control Centers, which contains records of all calls to PCCs in the United States. The Centers for Disease Control and Prevention monitors this database in real time for emerging issues or patterns of public health importance. For this study, all cases coded as a human “scorpion sting” (NPDS Generic Code 0205240) were obtained from the NPDS for 2010 to 2015.

Each record included information about the date and time of the report, location of exposure and caller, patient age and gender, signs and symptoms, management site, treatments, and outcome. Furthermore, for the 9 states with more than 100 annual reports (Alabama, Arizona, California, Florida, Georgia, New Mexico, Nevada, Oklahoma, and Texas), full 5-digit zip codes representing the caller location were also obtained. For the remaining states with fewer reported envenomations, only 3-digit zip codes were provided to ensure maximal individual de-identification.

Because zip codes are used by the US Postal Service (USPS) for mail delivery and are not meant to represent geographic areas, the USPS does not publish geographic boundaries for their active zip codes. To facilitate mapping, the US Census Bureau created Zip Code Tabulation Areas (ZCTAs) as a generalized geographic representation of zip codes. We obtained ZCTA boundary data, as well as 2016 versions of all other geographic boundaries and features used for mapping in this project, from the US Census Bureau (https://www.census.gov/geo/maps-data/data/tiger-line.html) as TIGER/Line shapefiles.

We obtained ZCTA population estimates for 2015 from the US Census Bureau’s American Fact Finder Download Center (https://factfinder.census.gov/faces/nav/jsf/pages/download_center.xhtml) by searching the 2015 American Community Survey 5-year estimates data set with a Geographic Type of 860 (“5-digit ZIP Code Tabulation Area”) and Table ID of B01003 (“Total Population”).

Inclusion and Exclusion Criteria

We obtained all cases coded as a human scorpion sting (NPDS Generic Code 0205240) from 2010 to 2015 from the NPDS. We excluded duplicate records (those having the same case number), calls from outside of the continental United States, and cases that recorded a confirmed nonexposure (i.e., no human envenomation) from analysis.

Some zip codes did not have a corresponding ZCTA because they either did not exist at the time of the 2010 ZCTA update, or they represented post office boxes or individual buildings that received high volumes of mail. We excluded these cases, together with records without a zip code, from mapping.

For the first objective of mapping exposure locations in the 9 states with more than 100 annual reports, we included only cases from these 9 states. Furthermore, only cases with the same exposure and caller locations were analyzed (e.g., “own residence” for both the exposure location and caller location). We did this to ensure the caller zip code represented the same location as the actual exposure. When calculating incidence data, we only included ZCTAs that reported an annual average of more than 5 calls during the study period for analysis to help protect privacy in low-incidence areas.

For the second objective of mapping cases suggestive of C. sculpturatus envenomation, we selected scorpion exposures that reported the unique effects of fasciculations or nystagmus from data across all 50 states. Although the clinical effects were more accurately described as involuntary motor activity and random eye movements (which can be dysconjugate), we chose the 2 effects that were the closest match available in NPDS. We included cases with different exposure and caller locations for this objective because we wished to capture all such reports regardless of location.

Data Analysis and Mapping

We converted 5-digit zip codes for each case to the corresponding ZCTA. We converted 3-digit zip codes for Utah to all corresponding 5-digit ZCTAs, except for those with only a single report. We then mapped exposure counts for each ZCTA using the GIS software ArcGIS Pro (version 1.4.1; Esri Inc, Redlands, California) and overlayed this on geographic boundary and feature data. We calculated incidence for each ZCTA for 2015, the final year of data, using the population estimates obtained from the US Census Bureau.

RESULTS

After we removed duplicate records (those having the same case number), calls from outside of the continental United States, and cases that recorded a confirmed nonexposure (i.e., no human envenomation), 105 230 nationwide scorpion exposure cases were available for the 6-year study period.

The 9 states of interest for the first objective accounted for 101 467 (96.4%) of these cases. When we selected only the cases for which the caller and exposure sites were the same, we identified 90 069 cases. Notably, 89 365 (99.2%) of these cases occurred in or around a residence. After we removed 3560 cases without a zip code and 5517 cases without a corresponding ZCTA, we used a total of 80 992 records for mapping (Table 1).

TABLE 1—

Scorpion Exposures Reported in 9 US States with More than 100 Exposures Reported per Year: 2010–2015

		Highest Case Counts for 2010–2015			Highest Incidence Per 100 000 Population in 2015
State	No. of Cases Mapped for 2010–2015	ZCTA	City	Count	ZCTA	City	Incidence
Arizona	57 168	85234	Gilbert	1 448	85045	Phoenix	677
		85086	Phoenix/Anthem	1 261	85718	Catalina Foothills	584
		85718	Catalina Foothills	1 221	85085	Phoenix	548
		85041	Phoenix	1 215	85263	Rio Verde	529
		85255	Scottsdale	1 200	85284	Phoenix	476
Texas	9 695	78245	San Antonio	100	79911	El Paso	213
		79936	El Paso	82	78070	Spring Branch	59
		78109	Converse	80	78261	San Antonio	46
Nevada	4 537	89074	Henderson	283	89138	Las Vegas	170
		89134	Las Vegas	192	89134	Las Vegas	147
		89122	Las Vegas	187	89135	Las Vegas	121
California	2 275	92252	Joshua Tree	36	92277	Twentynine Palms	33
		92065	Ramona	35	91302	Calabasas	28
		92277	Twentynine Palms	32	92262	Palm Springs	22
					92101	San Diego	22
Georgia	1 919	30016	Covington	62	30016	Covington	23
		30263	Newnan	35	30263	Newnan	11
		30252	McDonough	31
Oklahoma	1 825	73102	Oklahoma City	137	73102	Oklahoma City	209
		74120	Tulsa	58	74120	Tulsa	178
		73505	Lawton	43
New Mexico	1 401	88203	Roswell	111	87043	Rio Rancho	121
		87124	Rio Rancho	110	88201	Roswell	42
		88201	Roswell	94	88203	Roswell	42
Florida	1 266	32259	St Johns	29
		34472	Ocala	29
		32164	Palm Coast	25
Alabama	906	35242	Birmingham	39	35244	Birmingham	18
		35244	Birmingham	26
		35405	Tuscaloosa	17
Total	80 992

Note. The Zip Code Tabulation Areas (ZCTAs) with the most reported exposures from 2010–2015 and the greatest incidence of exposures in 2015 are listed for each state. Only ZCTAs with more than 5 calls in 2015 are listed.

The geographic distribution by ZCTA is shown in Figure 1. In general, low levels of exposures were broadly distributed throughout each of the 9 southern US states, with several notable exceptions. Table 1 lists ZCTAs with the highest case counts and incidences for each state.

FIGURE 1—

Total Scorpion Exposures by Zip Code Tabulation Area (ZCTA) for a 5-Year Period in 9 US States: 2010–2015

Note. The 9 states with more than 100 annual reports are included. Inset maps show cities with greater numbers of exposures with ZCTA boundaries (gray) and major roads (interstates, US routes, state routes in blue).

In the southeastern United States, there was an area of broad activity in the northern half of Alabama and Georgia, as well as a second area of activity that encompassed southern Georgia and Florida. These regions were separated by a distinct band of minimal activity that extended from southwestern Alabama northeast through central Georgia. The incidence rates and total calls were relatively low; Florida did not have any ZCTAs with more than 5 exposures in 2015.

In the southcentral United States, several areas had high numbers of scorpion exposures. In Oklahoma, 1 ZCTA in Oklahoma City (73102) and 1 in Tulsa (74120) had the largest numbers of exposures during the study period and the highest incidence in 2015 (209 and 178 per 100 000, respectively) in the state. These incidence rates were among the highest outside of Arizona. In Texas, reported exposures were greatest in areas around El Paso and San Antonio, extending northeast to Austin. Other scattered regions of increased activity in Texas included areas around Temple, as well as just west of Conroe and The Woodlands, which are both north of Houston. The highest incidence in Texas was a ZCTA in El Paso (79911), with 213 reports per 100 000 population in 2015. In New Mexico, areas around Roswell (southeastern part of the state) had high numbers, whereas a ZCTA north of Albuquerque (87043) had the greatest incidence in the state, at 121 per 100 000.

In the southwest, much of Arizona, particularly the Phoenix to Tucson corridor, Kingman in the northwest, Nogales in the south, and Safford in the east, had high frequencies. The greatest incidence was in southern Phoenix (85045), with 677 reports per 100 000 population. Arizona ZCTAs had the greatest incidence rates of any state in 2015 (Figure 2). In Nevada, activity was low, except around Las Vegas, which had incidences as high as 170.

FIGURE 2—

Arizona Scorpion Exposures per 100 000 Population by Zip Code Tabulation Area (ZCTA) in 2015

Note. Interstates and US routes are shown in blue. Inset map shows Phoenix with ZCTA boundaries (gray) and major roads (interstates, US routes, state routes in blue).

There were 1634 scorpion exposures that reported fasciculations and 2538 that recorded nystagmus. Zip codes were not reported for 64 of the fasciculation cases and 95 of the nystagmus cases (although the state was reported for these cases). Figure A (data available as a supplement to the online version of this article at http://www.ajph.org) maps the ZCTAs in and around Arizona that reported cases of fasciculations or nystagmus because this is the contiguous area where C. sculpturatus is known to inhabit. Cases from other states are included in the inset table in Figure A. Both clinical effects were reported throughout much of Arizona and southeastern Nevada, particularly Las Vegas. Interestingly, there were some cases of both fasciculations and nystagmus from Utah as well, where only 3-digit zip codes were available because this was not one of the states with more than 100 annual calls. Fasciculations were reported from zip codes 840 and 846, whereas nystagmus was reported from zip codes 840, 841, 845, and 847. For Utah, Figure A shows only those areas with more than 1 case from each 3-digit zip code. Southern California reported single cases of fasciculations in 3 ZCTAs (92024, 92252, and 92284) and no cases of nystagmus. New Mexico had single reports of fasciculations in 3 ZCTAs (87901, 88012, and 88201) and nystagmus in 2 ZCTAs (88061 and 87122). Western Texas (El Paso) had a single report of fasciculations (79932) and 1 report of nystagmus (79905).

DISCUSSION

Geographic analysis of scorpion exposures indicated their occurrence at baseline rates throughout many of the southern US states. The notable exceptions with greater incidences included large parts of Arizona and ZCTAs around several urban areas, including Las Vegas, Albuquerque, El Paso, San Antonio, Oklahoma City, and Tulsa.

The high frequencies in Texas around El Paso, San Antonio, and Austin remained consistent with Texas Poison Center Network data from 1998 to 2003,⁶ but our study identified several other areas with high activity in eastern Texas, just north of Houston. The single ZCTAs with high frequencies in Oklahoma City and in Tulsa were unusual in that each was considerably greater than the background frequencies in neighboring ZCTAs, with incidences among the highest seen outside of Arizona. In the southeastern United States, there appeared to be 2 separate areas of scorpion activity separated by a band across southern Alabama and central Georgia. This geographic distribution of exposure reports bore a striking resemblance to the published distributions of Vaejovis carolinianus (southern unstriped scorpion) in northern Alabama and Georgia (extending north into Kentucky) and the various Centruroides species found primarily in Florida.⁷

Mapping based on 2 clinical effects unique to severe C. sculpturatus envenomation indicated activity was concentrated in Arizona but also extended into Nevada (Las Vegas) and Utah. The degree of activity in Nevada appeared to be greater than suggested previously.² Precise mapping in Utah was limited by the availability of only 3-digit zip codes, but might be greater in the southern part of the state. Single scattered reports were also recorded in southern California, New Mexico, and western Texas, which agreed with the previously reported range of this species.

The few reports from other noncontiguous states might be related to exported cases or non-native scorpion species. This information might assist hospitals to assess the need to stock antivenom, and remind health care workers in these areas of the possibility of envenomation by this scorpion species.

Knowledge of scorpion exposure distribution could allow targeted education and prevention efforts in the affected areas. Even in Arizona, where the baseline frequency of exposures was considerably greater than elsewhere, there were still several ZCTAs with increased reports that could be targeted.

Previous studies combined PCC data with GIS to map issues relevant to public health, but these were limited to single states or provinces.^8–10 One previous mapping study of scorpion exposures was limited to the Phoenix area.¹¹ Our study used 5-digit zip code data from multiple states to display exposure activity at the level of individual ZCTAs. Despite NPDS data containing the zip code of only the caller, we were able to generate usable maps by limiting cases to those where the caller and exposure locations were the same. Such strategies could be used to generate maps useful for public education.

Limitations

NPDS data had several major limitations. First, data collection was passive and required callers to report an exposure to the PCC. Because there is no mandatory reporting of scorpion envenomation, the numbers reported were likely an underestimate, especially for more minor stings. The validity of any exposures that are reported depends on the caller providing correct information and the PCC staff accurately entering data. The identity of a scorpion species was not definitively known in most cases.

Interpretation of zip code data was limited by several factors. As described earlier, zip codes are used by the USPS to represent mail delivery routes and are not true geographic boundaries, nor are they constant. Mapping them to US Census Bureau ZCTAs was not perfect because there were no representations for zip codes corresponding to post office boxes or high-volume mail customers. Because we could not map some records due to missing data or lack of a corresponding ZCTA, the generated maps are a partial snapshot. For the same reason, incidences based on ZCTA must be interpreted with caution.

We hypothesized that the zip codes reported by callers corresponded to the location of the caller. Because most calls originated from a residence, we believed this to be a reasonable assumption. However, for the small number of exposures (0.8%) reported from sites such as a public place, restaurant, or other, this assumption might not be valid.

Severe C. sculpturatus envenomation (grades 3 and 4) can include involuntary motor and eye movements. Among the clinical effects available in NPDS, we chose fasciculations and nystagmus as the closest descriptors that might be used to code these manifestations because these are what are typically used in the Arizona PCCs. In areas with few C. sculpturatus exposures, PCCs might not have inquired about these particular effects because of unfamiliarity, or PCCs might have coded them differently. These clinical effects would not capture milder envenomations, which might be more common, and even when these effects were present, they might not have been documented. Although the effects we chose represented envenomation severity by this species, there was no evidence to suggest geographic differences in severity distribution. Consequently, we believed that the geographic distribution of fasciculations and nystagmus should represent a reasonable estimate of the total range of this species. Lastly, because we used only human exposure reports, areas with little human activity would not be represented. All of the preceding would contribute to underestimating the range of this species.

Despite these limitations in capturing the range of scorpion sting cases, which likely resulted in a conservative estimate, these data are important to improve awareness of the patterns of scorpion exposures in the nation.

Public Health Implications

Combining NPDS data and GIS software provided a useful visual analysis of scorpion calls to US poison centers. Prevention and education efforts could be targeted to the areas of greater scorpion exposure frequencies identified in this study. Although clinical effects consistent with envenomation by the Arizona bark scorpion were most common in Arizona and southeastern Nevada (around the Las Vegas area), health care providers should be aware that such cases also occur in southern California, New Mexico, western Texas, and Utah. This information might be helpful in decisions to stock antivenom.