Advertisement

Regional differences in EpiPen prescriptions in the United States: The potential role of vitamin D

      Background

      The epidemiology of anaphylaxis is uncertain, especially its geographic distribution.

      Objective

      To address this deficit, we examined regional rates of EpiPen prescriptions in the United States.

      Methods

      EpiPen prescriptions in 2004 were obtained for all 50 states and Washington, DC, from NDCHealth, Pharmaceutical Audit Suite (Alpharetta, Ga). Data included the number of total filled prescriptions, including refills, and the actual number of EpiPens prescribed. Several data sets were used to obtain state-specific populations, as well as multiple demographic, health, and weather characteristics. State population was used to calculate the average number of prescriptions written per person.

      Results

      Overall, there were 1,511,534 EpiPen prescriptions filled during 2004. These prescriptions accounted for 2,495,188 EpiPens. On average, there were 5.71 EpiPens prescribed per 1000 persons. Massachusetts had the highest number of prescriptions per 1000 persons (11.8), whereas Hawaii had the lowest (2.7). In addition to state-to-state variation, there was an obvious regional difference: New England (Connecticut, Rhode Island, Massachusetts, Vermont, New Hampshire, Maine) had the highest values, with 8 to 12 EpiPen prescriptions per 1000 persons, whereas the southern states (between and including California and Mississippi) had only 3 prescriptions per 1000 persons. The New England finding persisted even when controlling for all available factors (eg, population demographic characteristics, number of health care providers, prescriptions for other medications).

      Conclusion

      A strong north-south gradient was observed for the prescription of EpiPens in the United States, with the highest rates found in New England.

      Clinical implications

      The regional differences in EpiPen prescribing may provide important etiologic clues (vitamin D status) and merit further investigation.

      Key words

      The epidemiology of anaphylaxis in the general population is uncertain. A review article by Neugut et al
      • Neugut A.I.
      • Ghatak A.T.
      • Miller R.L.
      Anaphylaxis in the United States: an investigation into its epidemiology.
      estimated the US prevalence of anaphylaxis to be anywhere between 1% and 15%. By contrast, a population-based study in Olmsted County, Minn, concluded that the prevalence of anaphylaxis was less than 1%.
      • Yocum M.W.
      • Butterfield J.H.
      • Klein J.S.
      • Volcheck G.W.
      • Schroeder D.R.
      • Silverstein M.D.
      Epidemiology of anaphylaxis in Olmsted County: a population-based study.
      More recent studies from the United States, Canada, and the United Kingdom yield prevalence estimates ranging from 0.3% to 0.95%.
      • Bohlke K.
      • Davis R.L.
      • DeStefano F.
      • Marcy S.M.
      • Braun M.M.
      • Thompson R.S.
      Epidemiology of anaphylaxis among children and adolescents enrolled in a health maintenance organization.
      • Simons F.E.
      • Peterson S.
      • Black C.D.
      Epinephrine dispensing patterns for an out-of-hospital population: a novel approach to studying the epidemiology of anaphylaxis.
      • Peng M.M.
      • Jick H.
      A population-based study of the incidence, cause, and severity of anaphylaxis in the United Kingdom.
      • Gupta R.
      • Sheikh A.
      • Strachan D.P.
      • Anderson H.R.
      Burden of allergic disease in the UK: secondary analyses of national databases.
      The study yielding the highest prevalence estimate (0.95%) was by Simons et al
      • Simons F.E.
      • Peterson S.
      • Black C.D.
      Epinephrine dispensing patterns for an out-of-hospital population: a novel approach to studying the epidemiology of anaphylaxis.
      and used a novel approach to estimating the burden of anaphylaxis: evaluating medication-dispensing data for epinephrine for out-of-hospital treatment during a 5-year period in Manitoba, Canada. A recent expert roundtable concluded that this approach may provide one of the best insights into the actual frequency of anaphylaxis.
      • Lieberman P.
      • Camargo Jr., C.A.
      • Bohlke K.
      • Jick H.
      • Miller R.L.
      • Sheikh A.
      • et al.
      Epidemiology of anaphylaxis: findings of the American College of Allergy, Asthma and Immunology (ACAAI) Epidemiology of Anaphylaxis Working Group.
      This approach also permits investigation of the geographic distribution of anaphylaxis, which is unknown. Our objective was to describe the geographic distribution of anaphylaxis in the United States and to explore potential mediators of any observed geographic differences. To do this, we examined state-specific EpiPen (Dey, Napa, Calif) prescriptions in 2004. Although our analysis was largely exploratory, we hypothesized—on the basis of the vitamin D–asthma hypothesis
      • Camargo Jr., C.A.
      • Rifas-Shiman S.L.
      • Litonjua A.A.
      • Rich-Edwards J.W.
      • Weiss S.T.
      • Gold D.R.
      • et al.
      Maternal intake of vitamin D during pregnancy and risk of recurrent wheeze in children at 3 y of age.
      —that northern states with less year-round sunlight (eg, New England) would have higher EpiPen prescriptions than southern states.

      Methods

      US regions and divisions were defined based on the US Census Bureau classifications.

      US Census Bureau web site. Available at: http://www.census.gov/population/www/index.html. Accessed August 1, 2006.

      The 9 divisions of the United States were New England (Maine, New Hampshire, Vermont, Massachusetts, Connecticut, Rhode Island), Middle Atlantic (New York, Pennsylvania, New Jersey), South Atlantic (Delaware, Washington, DC, Maryland, West Virginia, Virginia, North Carolina, South Carolina, Georgia, Florida), East North Central (Michigan, Wisconsin, Illinois, Indiana, Ohio), East South Central (Kentucky, Tennessee, Alabama, Mississippi), West North Central (North Dakota, Minnesota, South Dakota, Nebraska, Iowa, Kansas, Missouri), West South Central (Oklahoma, Arkansas, Texas, Louisiana), Mountain (Montana, Idaho, Wyoming, Nevada, Utah, Colorado, Arizona, New Mexico), and Pacific (Washington, Oregon, California, Alaska, Hawaii). The 4 US regions were Northeast (New England and Middle Atlantic), Midwest (East North Central and West North Central), South (South Atlantic, East South Central, and West South Central), and West (Mountain and Pacific). Finally, region was divided into 2 groups—New England versus other US regions—to evaluate further the most clear-cut geographic difference in EpiPen prescribing.
      EpiPen prescriptions and other medication prescriptions during 2004, for all 50 US states and District of Columbia, were obtained from NDCHealth, Pharmaceutical Audit Suite (Alpharetta, Ga). In 2004, there were no competitive self-injectable epinephrine products in the US marketplace. Data were obtained on the number of total filled prescriptions (including refills) and the actual number of EpiPens dispensed for EpiPen 0.3 mg, EpiPen Jr, and total EpiPens (0.3 mg and Jr combined). Additional data obtained from NDCHealth were the number of allergist, pediatrician, adult primary care (ie, adult generalist, family physician), and emergency physician prescription writers per state, and filled prescriptions for a few other medications (eg, inhaled corticosteroids, atorvastatin). Data on demographic factors (eg, age, race/ethnicity, income) and state population were downloaded from the US Census.

      US Census Bureau web site. Available at: http://www.census.gov/population/www/index.html. Accessed August 1, 2006.

      Data on other health characteristics were obtained from state-specific estimates provided by the 2004 Behavioral Risk Factor Surveillance System website.

      Centers for Disease Control and Prevention. Behavioral Risk Factor Surveillance System (BRFSS) web site. Available at: http://www.cdc.gov/asthma/brfss/default.htm. Accessed June 1, 2006.

      Health characteristics data included health insurance status, history of asthma, history of diabetes, self-reported general health status, smoking history, alcohol consumption, and flu and pneumonia vaccinations for individuals age 65 years and older. State-level data on average annual temperature, average annual precipitation, average January normal temperature, and average July normal temperature were obtained from the National Oceanic and Atmospheric Administration for 1971 to 2000.

      National Oceanic and Atmospheric Adminstration (NOAA) web site. Available at: http://www.noaa.gov/. Accessed June 1, 2006.

      Estimated 2004 melanoma incidence, an admittedly crude surrogate for sun exposure,
      • Ivry G.B.
      • Ogle C.A.
      • Shim E.K.
      Role of sun exposure in melanoma.
      • Oliveria S.A.
      • Saraiya M.
      • Geller A.C.
      • Heneghan M.K.
      • Jorgensen C.
      Sun exposure and risk of melanoma.
      was obtained from the American Cancer Society.
      • Jemal A.
      • Tiwari R.C.
      • Murray T.
      • Ghafoor A.
      • Samuels A.
      • Ward E.
      • et al.
      Cancer statistics, 2004.

       Statistical analysis

      All analyses were performed using STATA 9.0 (StataCorp, College Station, Tex). State population was used to calculate the number of prescriptions filled per person. All rates are presented per 1000 state population. Means are presented with SDs and medians with interquartile ranges. The association between factors of interest and region (ie, New England vs other US regions) was evaluated using the Student t test and Kruskal-Wallis test, as appropriate. Multivariate linear regression was used to evaluate the association between state factors and EpiPen prescriptions. All β coefficients are presented with 95% CIs.

      Results

      In 2004, including refills, there were a total of 1,511,534 EpiPen 0.3 mg and EpiPen Jr prescriptions filled, with EpiPen 0.3 mg accounting for 78% (1,178,229) of prescriptions and EpiPen Jr 22% (333,305). These 1.5 million EpiPen prescriptions led to dispensing of 2,495,188 EpiPens or an average of 1.6 EpiPens per prescription. There were an average of 1.6 EpiPen 0.3 mg devices per prescription and an average of 1.9 EpiPen Jr devices per prescription.
      We found considerable state-to-state variation when looking at the absolute number of EpiPens prescribed, from a high of 125,540 in New York to a low of 2369 in Wyoming. When adjusting for population differences, by looking at the prescription rates per 1000 state population, there were an average of 5.71 EpiPens prescribed per 1000 persons. As shown in Fig 1, state-to-state variation persisted, with the highest rates seen in Massachusetts (11.8) and the lowest rates seen in Hawaii (2.7).
      Figure thumbnail gr1
      Fig 1Number of EpiPen prescriptions per 1000 persons by state, ranging from 2.7 (Hawaii) to 11.8 (Massachusetts).
      In addition to state-to-state variation, we observed a striking regional difference (Fig 2). The Northeast had the most EpiPen prescriptions, with 8 to 12 prescriptions per 1000 persons, and the Southern states had ¼ the prescriptions at 2 to 3 prescriptions per 1000 persons. When separating EpiPen 0.3 mg and EpiPen Jr, we found that these regional trends were similar with both sizes of EpiPen (data not shown). In an unadjusted linear regression model, all other US regions had significantly fewer EpiPen prescriptions than New England (reference group): Mid Atlantic (β, –3.64; 95% CI, –5.46 to –1.84), East North Central (β, –4.96; 95% CI, –6.50 to –3.41), West North Central (β, –5.84; 95% CI, –7.27 to –4.42), South Atlantic (β, –4.29; 95% CI, –5.64 to –2.94), East South Central (β, –5.60; 95% CI, –7.25 to –3.95), West South Central (β, –6.82; 95% CI, –8.47 to –5.17), Mountain (β, –6.35; 95% CI, –7.73 to –4.97), and Pacific (β, –5.40; 95% CI, –6.95 to –3.85).
      Figure thumbnail gr2
      Fig 2Regional differences in EpiPen prescriptions per 1000 persons.
      Table I shows population demographic and health characteristics in the New England region compared with those in all other US regions combined. The state populations in the New England region had a higher median age, median household income, and percentage of people reporting health insurance. The number of allergists and emergency physicians was similar, but New England also had more pediatricians and adult primary care providers (per 1000 population). New England also prescribed more inhaled corticosteroids and atorvastatin per 1000 population; other studied medications did not significantly differ.
      Table IState characteristics in New England region versus other US regions
      Data presented as means ± SDs, or medians with interquartile ranges.
      New England regionOther US regionsP value
      Median age in state (y)39 ± 136 ± 2.002
      Percent female in state51 ± 0.451 ± 0.8.07
      Percent white in state92 ± 580 ± 14.05
      Median household income (US $)62,696 ± 857652,500 ± 7688.004
      Percent high school graduates84 ± 382 ± 4.14
      Percent any type of health insurance coverage89 ± 284 ± 4.01
      No. of healthcare providers (per 1000 population)
       Allergists0.01 (0.01-0.02)0.01 (0.01-0.02).95
       Pediatricians0.32 (0.26-0.36)0.22 (0.17-0.26).006
       Adult primary care providers1.31 (1.14-1.47)1.02 (0.90-1.12).002
       Emergency physicians0.17 (0.14-0.22)0.14 (0.13-0.17).08
      No. of allergy/asthma medications (per 1000 population)
       Albuterol31.2 (23.9-34.6)36.2 (28.3-40.8).13
       Inhaled corticosteroids62.3 (56.5-62.4)37.1 (30.7-41.5)<.001
       Montelukast70.3 (62.1-83.2)79.3 (62.3-86.5).77
       Pimecrolimus cream11.2 (8.2-15.7)12.4 (10.0-16.4).38
      No. of other medications (per 1000 population)
       Atorvastatin349.4 (303.9-453.8)226.5 (185.0-262.6).001
       Oral hypoglycemic agents
      Includes sulfonylureas, meglitinides, amino acid derivatives, biguanides, insulin sensitizers, α-glucosidase inhibitors, noninsulin combinations, and noninsulin others.
      319.7 (307.2-358.3)337.3 (280.0-391.1).60
       Hypertension medications
      Includes Universal System of Classification codes 41000 (Diuretics, Ethical) and 31000 (Vascular Agents).
      1957.4 (1905.7-2072.9)1978.5 (1577.0-2346.6).88
      Average temperature (°F)46 (43-49)53 (48-59).06
      Average precipitation (in)46 (43-48)39 (23-46).10
      Melanoma incidence280 (280-700)910 (420-1390).14
      Data presented as means ± SDs, or medians with interquartile ranges.
      Includes sulfonylureas, meglitinides, amino acid derivatives, biguanides, insulin sensitizers, α-glucosidase inhibitors, noninsulin combinations, and noninsulin others.
      Includes Universal System of Classification codes 41000 (Diuretics, Ethical) and 31000 (Vascular Agents).
      In univariate analyses, several state-level factors were associated with the number of EpiPen prescriptions per state (Table II). Number of EpiPen prescriptions was positively associated with percentage of the state population that was female but did not differ by other sociodemographic factors. Strong associations were seen for number of healthcare providers (especially allergists), and with medication prescriptions of all types. Average temperature and average precipitation were not significantly associated with number of EpiPen prescriptions. Melanoma incidence had a significant inverse association with EpiPen prescriptions.
      Table IIUnivariate predictors of number of EpiPen prescriptions per state
      β95% CIP value
      Median age (per ↑1 y)−95−3826, 3637.96
      Percent female in state (per ↑1%)1.260.48, 2.05.002
      Percent white in state (per ↑1%)−106−699, 486.72
      Median household income (per ↑$10,000)9005−407, 18418.06
      Percent high school graduates (per ↑1%)−1314−3152, 523.16
      Percent any type of health insurance coverage191−1865, 2249.85
      Healthcare providers (per ↑1 provider)
       Allergists233202, 265<.001
       Pediatricians1513, 17<.001
       Adult primary care providers4.03.7, 4.6<.001
       Emergency physicians3025, 35<.001
      Allergy/asthma medications (per ↑1 prescription)
       Albuterol0.100.09, 0.12<.001
       Inhaled corticosteroids0.120.10, 0.13<.001
       Montelukast0.060.05, 0.07<.001
       Pimecrolimus cream0.290.24, 0.33<.001
      Other medications (per ↑1 prescription)
       Atorvastatin0.0190.017, 0.021<.001
       Oral hypoglycemic agents
      Includes sulfonylureas, meglitinides, amino acid derivatives, biguanides, insulin sensitizers, α-glucosidase inhibitors, noninsulin combinations, and noninsulin others.
      0.0120.011, 0.014<.001
       Hypertension medications
      Includes USC codes 41000 (Diuretics, Ethical) and 31000 (Vascular Agents).
      0.0020.002, 0.003<.001
      Average temperature (per ↑1°F)472−468, 1412.32
      Average precipitation (per ↑1 inch)371−186, 928.19
      Melanoma incidence2420, 27<.001
      Includes sulfonylureas, meglitinides, amino acid derivatives, biguanides, insulin sensitizers, α-glucosidase inhibitors, noninsulin combinations, and noninsulin others.
      Includes USC codes 41000 (Diuretics, Ethical) and 31000 (Vascular Agents).
      Other factors describing the general health of the state population also were evaluated, including history of asthma, history of diabetes, self-reported general health status, cigarette smoking, alcohol consumption, and flu and pneumonia vaccinations for individuals age 65 years and older. We also examined confidential data on Dey marketing efforts in the different states. However, none of these factors were significantly associated with the number of EpiPen prescriptions filled (data not shown).
      To identify population characteristics that might mediate the marked regional differences, we evaluated the rate of EpiPen prescriptions in the New England (highest prescription rates) compared with all other regions while controlling for all available factors (Table III). Controlling for these factors attenuated the higher rate of EpiPen prescriptions in the New England region, but the finding remained statistically significant (P < .001).
      Table IIIMultivariate model of regional EpiPen prescriptions (per 1000 state population), comparing New England region vs other US regions
      β95% CIP value
      Unadjusted5.394.11-6.67<.001
      Model 1
      Model 1 controls for 6 factors: median age, percent female, percent white, median household income, percent high school graduates, and percent with health insurance.
      4.072.77-5.36<.001
      Model 2
      Model 2 controls for above, plus 4 factors: number of allergists, pediatricians, adult primary care providers, and emergency physicians.
      3.792.20-5.39<.001
      Model 3
      Model 3 controls for above, plus 7 factors: number of prescriptions for albuterol, inhaled corticosteroids, montelukast, pimecrolimus cream, atorvastatin, oral hypoglycemic agents, and hypertension medications.
      3.591.94-5.23<.001
      Model 4
      Model 4 controls for above, plus melanoma incidence.
      3.551.86-5.24<.001
      Model 1 controls for 6 factors: median age, percent female, percent white, median household income, percent high school graduates, and percent with health insurance.
      Model 2 controls for above, plus 4 factors: number of allergists, pediatricians, adult primary care providers, and emergency physicians.
      Model 3 controls for above, plus 7 factors: number of prescriptions for albuterol, inhaled corticosteroids, montelukast, pimecrolimus cream, atorvastatin, oral hypoglycemic agents, and hypertension medications.
      § Model 4 controls for above, plus melanoma incidence.

      Discussion

      To address uncertainty about the geographic distribution of anaphylaxis and overcome sparse and often problematic data on healthcare utilization for this condition,
      • Lieberman P.
      • Camargo Jr., C.A.
      • Bohlke K.
      • Jick H.
      • Miller R.L.
      • Sheikh A.
      • et al.
      Epidemiology of anaphylaxis: findings of the American College of Allergy, Asthma and Immunology (ACAAI) Epidemiology of Anaphylaxis Working Group.
      we examined EpiPen prescription rates across the United States. We found a strong north-south gradient, with the highest rates of EpiPen prescriptions in New England. Population-adjusted rates (ie, number of EpiPen prescriptions filled per 1000 people) were positively associated with several different factors, including percentage of the population that was female, the number of healthcare providers (especially allergists) per 1000 people, and the frequency of medication prescribing in general. Nevertheless, a multivariate analysis controlling for all of these factors suggested that these factors did not mediate the observed north-south gradient.
      Ecologic studies are, by their very nature, exploratory, but they can yield important etiologic clues about disease. For example, it has been known for decades that the incidence of multiple sclerosis varies by latitude.
      • Hernan M.A.
      • Olek M.J.
      • Ascherio A.
      Geographic variation of MS incidence in two prospective studies of US women.
      This geographic pattern may be explained by differences in vitamin D status, with higher incidence of multiple sclerosis at higher latitudes where vitamin D insufficiency is more common.
      • Brown S.J.
      The role of vitamin D in multiple sclerosis.
      Vitamin D3 (cholecalciferol) is a vital nutrient available from dietary sources (eg, fortified milk, nutritional supplements), but most is made in the skin after direct exposure to sunlight.
      • Feldman D.
      • Pike J.W.
      • Glorieux F.H.
      Vitamin D.
      It has become clear, especially in the northeastern United States,
      • Webb A.R.
      • Kline L.
      • Holick M.F.
      Influence of season and latitude on the cutaneous synthesis of vitamin D3: exposure to winter sunlight in Boston and Edmonton will not promote vitamin D3 synthesis in human skin.
      • Thomas M.K.
      • Lloyd-Jones D.M.
      • Thadhani R.I.
      • Shaw A.C.
      • Deraska D.J.
      • Kitch B.T.
      • et al.
      Hypovitaminosis D in medical inpatients.
      that a large proportion of Americans have inadequate vitamin D intake, as reflected by serum 25(OH)D levels.
      • Hollis B.W.
      Circulating 25-hydroxyvitamin D levels indicative of vitamin D sufficiency: implications for establishing a new effective dietary intake recommendation for vitamin D.
      On the basis of these data, and emerging evidence of an inverse association between vitamin D status and risk of childhood wheezing,
      • Camargo Jr., C.A.
      • Rifas-Shiman S.L.
      • Litonjua A.A.
      • Rich-Edwards J.W.
      • Weiss S.T.
      • Gold D.R.
      • et al.
      Maternal intake of vitamin D during pregnancy and risk of recurrent wheeze in children at 3 y of age.
      we hypothesized a priori that we would find a north-south gradient in anaphylaxis, as reflected by regional differences in filled EpiPen prescriptions. Indeed, residence in the northeastern United States was the strongest independent predictor of EpiPen prescriptions per 1000 persons.
      Although measurement of serum 25(OH)D levels is the best available approach to determine vitamin D status,
      • Hollis B.W.
      Circulating 25-hydroxyvitamin D levels indicative of vitamin D sufficiency: implications for establishing a new effective dietary intake recommendation for vitamin D.
      average state-specific or region-specific values are not available. Indeed, the measurement of sun exposure itself has proven quite challenging. Epidemiologists who study the relation of sun exposure with skin cancer continue to debate the advantages and disadvantages of self-reported information on regular versus intermittent sun exposure, frequency of sun burns, regularity of sunscreen use, and more.
      • Ivry G.B.
      • Ogle C.A.
      • Shim E.K.
      Role of sun exposure in melanoma.
      • Oliveria S.A.
      • Saraiya M.
      • Geller A.C.
      • Heneghan M.K.
      • Jorgensen C.
      Sun exposure and risk of melanoma.
      For lack of any better available alternative, we used melanoma incidence rates as an admittedly crude surrogate measure of sun exposure and the resulting vitamin D levels. Sun exposure generally has a positive association with melanoma risk, but there is some evidence that chronic (regular) sun exposure may actually reduce risk of melanoma.
      • Ivry G.B.
      • Ogle C.A.
      • Shim E.K.
      Role of sun exposure in melanoma.
      • Oliveria S.A.
      • Saraiya M.
      • Geller A.C.
      • Heneghan M.K.
      • Jorgensen C.
      Sun exposure and risk of melanoma.
      Thus, we recognize the probable limitations of our exploratory analysis. Regardless, we found that melanoma incidence rates had a significant inverse association with EpiPen prescriptions, but that adjustment for melanoma incidence did not eliminate the observed north-south gradient. Other meteorology variables, such as average temperature and precipitation, were not associated with the number of EpiPen prescriptions, but these variables are even more indirect measures of sunlight exposure, let alone vitamin D status.
      Laboratory research suggests several potential mechanisms for how vitamin D could affect risk of allergic reactions and anaphylaxis. For example, vitamin D modulates antigen-presenting cells such as macrophages,
      • Griffin M.D.
      • Xing N.
      • Kumar R.
      Vitamin D and its analogs as regulators of immune activation and antigen presentation.
      • Lin R.
      • White J.H.
      The pleiotropic actions of vitamin D.
      as well as the generation of regulatory T cells
      • Gregori S.
      • Giarratana N.
      • Smiroldo S.
      • Uskokovic M.
      • Adorini L.
      A 1alpha,25-dihydroxyvitamin D(3) analog enhances regulatory T-cells and arrests autoimmune diabetes in NOD mice.
      • Meehan M.A.
      • Kerman R.H.
      • Lemire J.M.
      1,25-Dihydroxyvitamin D3 enhances the generation of nonspecific suppressor cells while inhibiting the induction of cytotoxic cells in a human MLR.
      that express potentially inhibitory cytokines (IL-10 and TGF-β), and the ability to inhibit antigen-specific T-cell activation potently.
      • Schwartz R.H.
      Natural regulatory T cells and self-tolerance.
      Although many laboratory studies suggest that vitamin D induces a shift in the balance between TH1 and TH2-type cytokines toward TH2 dominance,
      • Cantorna M.T.
      • Zhu Y.
      • Froicu M.
      • Wittke A.
      Vitamin D status, 1,25-dihydroxyvitamin D3, and the immune system.
      Pichler et al
      • Pichler J.
      • Gerstmayr M.
      • Szepfalusi Z.
      • Urbanek R.
      • Peterlik M.
      • Willheim M.
      1 Alpha,25(OH)2D3 inhibits not only Th1 but also Th2 differentiation in human cord blood T cells.
      found that in CD4+ as well as CD8+ human cord blood cells, vitamin D not only inhibits IL-12–generated IFN-γ production but also suppresses IL-4 and IL-4–induced expression of IL-13. Thus, the differences between the studies on the TH1-TH2 dominance may lie in the timing of exposure of the cells to vitamin D (ie, prenatal versus postnatal); the response of naive T cells to vitamin D exposure may differ from that of mature cells when exposed to vitamin D.
      • Annesi-Maesano I.
      Perinatal events, vitamin D, and the development of allergy.
      Another possibility is that the association differs by dose. In other words, nonpharmacologic doses of vitamin D (eg, as obtained from sunlight exposure) may have different consequences than high-dose oral supplementation, where an excess of vitamin D may indeed have opposite effects. These hypotheses merit further investigation.
      Our exploratory study has several potential limitations. First, we assume that EpiPen prescriptions accurately reflect the underlying prevalence of anaphylaxis (ie, that geographic locations with higher EpiPen prescriptions have more anaphylaxis). Although the exact nature of the association between EpiPen prescriptions and actual anaphylaxis prevalence is not clear, this approach may be preferable to other more problematic sources of anaphylaxis data.
      • Simons F.E.
      • Peterson S.
      • Black C.D.
      Epinephrine dispensing patterns for an out-of-hospital population: a novel approach to studying the epidemiology of anaphylaxis.
      Accordingly, research on EpiPen prescriptions has been promoted by leading authorities in anaphylaxis epidemiology.
      • Lieberman P.
      • Camargo Jr., C.A.
      • Bohlke K.
      • Jick H.
      • Miller R.L.
      • Sheikh A.
      • et al.
      Epidemiology of anaphylaxis: findings of the American College of Allergy, Asthma and Immunology (ACAAI) Epidemiology of Anaphylaxis Working Group.
      Second, we recognize that there are apparent inconsistencies in the north-south gradient (eg, Why doesn't Alaska have the highest rate?) and there is a suggestion of an east-west gradient across the United States. These inconsistencies are not surprising in an ecologic analysis of US states and, in this case, may reflect differences in dispensing of epinephrine ampules and/or epinephrine-containing inhalers, practice variation without medical explanation, or myriad other factors. Future studies will need to explore these possibilities in more detail. Third, our study may suffer from the ecologic fallacy, in which individual exposures are not linked with individual outcomes and, therefore, one cannot be certain of the actual associations in individual patients. In our study, we can be confident about the observed regional differences in EpiPen prescriptions but are less confident that the multivariate analysis adequately controlled for the potential confounding factors (eg, socioeconomic factors). For this reason, we encourage additional research to confirm the observed north-south gradient and to uncover the factors that best explain it.

       Summary

      We observed a surprisingly strong north-south gradient for the prescription of EpiPens in the United States. The southwestern United States (and Hawaii), with its warmer climate and more year-round sunlight, had the least EpiPen prescriptions per 1000 persons, whereas those in the northern United States (especially New England) had the highest rates. The regional differences may result from several different associations, but we were not able to eliminate the finding by adjusting for regional differences in many sociodemographic and healthcare delivery factors. We suggest that these data provide additional support for the hypothesized link between low vitamin D levels and respiratory/allergic disorders.
      • Camargo Jr., C.A.
      • Rifas-Shiman S.L.
      • Litonjua A.A.
      • Rich-Edwards J.W.
      • Weiss S.T.
      • Gold D.R.
      • et al.
      Maternal intake of vitamin D during pregnancy and risk of recurrent wheeze in children at 3 y of age.
      Given the paucity of data on anaphylaxis risk factors
      • Lieberman P.
      • Camargo Jr., C.A.
      • Bohlke K.
      • Jick H.
      • Miller R.L.
      • Sheikh A.
      • et al.
      Epidemiology of anaphylaxis: findings of the American College of Allergy, Asthma and Immunology (ACAAI) Epidemiology of Anaphylaxis Working Group.
      and ongoing concerns about magnitude of this allergic problem,
      • Simons F.E.
      Anaphylaxis, killer allergy: long-term management in the community.
      we believe this novel hypothesis merits further study.

      References

        • Neugut A.I.
        • Ghatak A.T.
        • Miller R.L.
        Anaphylaxis in the United States: an investigation into its epidemiology.
        Arch Intern Med. 2001; 161: 15-21
        • Yocum M.W.
        • Butterfield J.H.
        • Klein J.S.
        • Volcheck G.W.
        • Schroeder D.R.
        • Silverstein M.D.
        Epidemiology of anaphylaxis in Olmsted County: a population-based study.
        J Allergy Clin Immunol. 1999; 104: 452-456
        • Bohlke K.
        • Davis R.L.
        • DeStefano F.
        • Marcy S.M.
        • Braun M.M.
        • Thompson R.S.
        Epidemiology of anaphylaxis among children and adolescents enrolled in a health maintenance organization.
        J Allergy Clin Immunol. 2004; 113: 536-542
        • Simons F.E.
        • Peterson S.
        • Black C.D.
        Epinephrine dispensing patterns for an out-of-hospital population: a novel approach to studying the epidemiology of anaphylaxis.
        J Allergy Clin Immunol. 2002; 110: 647-651
        • Peng M.M.
        • Jick H.
        A population-based study of the incidence, cause, and severity of anaphylaxis in the United Kingdom.
        Arch Int Med. 2004; 164: 317-319
        • Gupta R.
        • Sheikh A.
        • Strachan D.P.
        • Anderson H.R.
        Burden of allergic disease in the UK: secondary analyses of national databases.
        Clin Exp Allergy. 2004; 34: 520-526
        • Lieberman P.
        • Camargo Jr., C.A.
        • Bohlke K.
        • Jick H.
        • Miller R.L.
        • Sheikh A.
        • et al.
        Epidemiology of anaphylaxis: findings of the American College of Allergy, Asthma and Immunology (ACAAI) Epidemiology of Anaphylaxis Working Group.
        Ann Allergy Asthma Immunol. 2006; 97: 596-602
        • Camargo Jr., C.A.
        • Rifas-Shiman S.L.
        • Litonjua A.A.
        • Rich-Edwards J.W.
        • Weiss S.T.
        • Gold D.R.
        • et al.
        Maternal intake of vitamin D during pregnancy and risk of recurrent wheeze in children at 3 y of age.
        Am J Clin Nutr. 2007; 85: 788-795
      1. US Census Bureau web site. Available at: http://www.census.gov/population/www/index.html. Accessed August 1, 2006.

      2. Centers for Disease Control and Prevention. Behavioral Risk Factor Surveillance System (BRFSS) web site. Available at: http://www.cdc.gov/asthma/brfss/default.htm. Accessed June 1, 2006.

      3. National Oceanic and Atmospheric Adminstration (NOAA) web site. Available at: http://www.noaa.gov/. Accessed June 1, 2006.

        • Ivry G.B.
        • Ogle C.A.
        • Shim E.K.
        Role of sun exposure in melanoma.
        Dermatol Surg. 2006; 32: 481-492
        • Oliveria S.A.
        • Saraiya M.
        • Geller A.C.
        • Heneghan M.K.
        • Jorgensen C.
        Sun exposure and risk of melanoma.
        Arch Dis Child. 2006; 91: 131-138
        • Jemal A.
        • Tiwari R.C.
        • Murray T.
        • Ghafoor A.
        • Samuels A.
        • Ward E.
        • et al.
        Cancer statistics, 2004.
        CA Cancer J Clin. 2004; 54: 8-29
        • Hernan M.A.
        • Olek M.J.
        • Ascherio A.
        Geographic variation of MS incidence in two prospective studies of US women.
        Neurology. 1999; 53: 1711-1718
        • Brown S.J.
        The role of vitamin D in multiple sclerosis.
        Ann Pharmacother. 2006; 40: 1158-1161
        • Feldman D.
        • Pike J.W.
        • Glorieux F.H.
        Vitamin D.
        2nd ed. Elsevier Academic Press, Amsterdam2005
        • Webb A.R.
        • Kline L.
        • Holick M.F.
        Influence of season and latitude on the cutaneous synthesis of vitamin D3: exposure to winter sunlight in Boston and Edmonton will not promote vitamin D3 synthesis in human skin.
        J Clin Endocrinol Metab. 1988; 67: 373-378
        • Thomas M.K.
        • Lloyd-Jones D.M.
        • Thadhani R.I.
        • Shaw A.C.
        • Deraska D.J.
        • Kitch B.T.
        • et al.
        Hypovitaminosis D in medical inpatients.
        N Engl J Med. 1998; 338: 777-783
        • Hollis B.W.
        Circulating 25-hydroxyvitamin D levels indicative of vitamin D sufficiency: implications for establishing a new effective dietary intake recommendation for vitamin D.
        J Nutr. 2005; 135: 317-322
        • Griffin M.D.
        • Xing N.
        • Kumar R.
        Vitamin D and its analogs as regulators of immune activation and antigen presentation.
        Annu Rev Nutr. 2003; 23: 117-145
        • Lin R.
        • White J.H.
        The pleiotropic actions of vitamin D.
        Bioessays. 2004; 26: 21-28
        • Gregori S.
        • Giarratana N.
        • Smiroldo S.
        • Uskokovic M.
        • Adorini L.
        A 1alpha,25-dihydroxyvitamin D(3) analog enhances regulatory T-cells and arrests autoimmune diabetes in NOD mice.
        Diabetes. 2002; 51: 1367-1374
        • Meehan M.A.
        • Kerman R.H.
        • Lemire J.M.
        1,25-Dihydroxyvitamin D3 enhances the generation of nonspecific suppressor cells while inhibiting the induction of cytotoxic cells in a human MLR.
        Cell Immunol. 1992; 140: 400-409
        • Schwartz R.H.
        Natural regulatory T cells and self-tolerance.
        Nat Immunol. 2005; 6: 327-330
        • Cantorna M.T.
        • Zhu Y.
        • Froicu M.
        • Wittke A.
        Vitamin D status, 1,25-dihydroxyvitamin D3, and the immune system.
        Am J Clin Nutr. 2004; 80: 1717S-1720S
        • Pichler J.
        • Gerstmayr M.
        • Szepfalusi Z.
        • Urbanek R.
        • Peterlik M.
        • Willheim M.
        1 Alpha,25(OH)2D3 inhibits not only Th1 but also Th2 differentiation in human cord blood T cells.
        Pediatr Res. 2002; 52: 12-18
        • Annesi-Maesano I.
        Perinatal events, vitamin D, and the development of allergy.
        Pediatr Res. 2002; 52: 3-5
        • Simons F.E.
        Anaphylaxis, killer allergy: long-term management in the community.
        J Allergy Clin Immunol. 2006; 117: 367-377