Patients with diabetic ketoacidosis (DKA) are often managed in the emergency department before hospital admission. DKA hospitalizations comprise a significant portion of health care costs for diabetes (1). Although mortality for DKA has fallen, it remains an important cause of diabetes-associated death, especially among younger patients with diabetes (2). Prior analyses of DKA have been single-center intensive care unit (ICU) studies or based on hospital discharges (3–5). Patients may, however, be treated in the emergency department and then admitted to a non-ICU setting or discharged; the frequency of these practices is not known. We sought to describe the epidemiology of emergency department visits with DKA.

We analyzed the emergency department component of the 1993–2003 U.S. National Hospital Ambulatory Medical Care Survey (NHAMCS). Our institutional review board waived review of this analysis. Methodological details are described elsewhere (6–8). Briefly, NHAMCS uses a four-stage sampling strategy covering geographic primary sampling units, hospitals within primary sampling units, emergency departments within hospitals, and patients within emergency departments. Hospitals were stratified by region, presence of emergency department, ownership type, and size. Within each stratum, hospitals were selected with a probability proportional to the number of emergency department visits. Data were collected during randomly assigned 4-week periods. Data forms include demographic information, emergency department disposition (i.e., admission, transfer, and discharge), and up to three ICD-9 discharge diagnoses. For the present analysis, we identified DKA visits based on ICD-9 code 250.1x, the unique code for DKA, in any of the diagnosis fields. This methodology has been utilized in prior Centers for Disease Control (CDC) analyses (3,9).

We used Stata 9.0 software (StatCorp, College Station, TX) to determine nationally representative estimates and 95% CIs using assigned patient weights, which adjust for selection probability. Where the relative SE was >30% or estimates were based on <30 cases, data are reported only as the number of observations. We calculated rates based on U.S. Census Bureau, CDC, and NHAMCS data.

Between 1993 and 2003, DKA accounted for ∼753,000 visits (95% CI 610,000–895,000) or an average 68,000 visits/year.

Table 1 shows patient and hospital characteristics of emergency department visits with DKA. Most DKA visits were evenly distributed among patients aged 10–50 years. The large majority of DKA patients (87% [95% CI 81–92]) were admitted, with most admissions to a non-ICU setting. DKA visit rates per 10,000 U.S. population were higher for black than white patients. Rates per 10,000 U.S. population and per 10,000 emergency department visits were otherwise similar across all demographic factors.

The rate of emergency department visits for DKA per 10,000 U.S. population with diabetes was 64 (95% CI 52–76). Among this population, there was no significant difference in visit rates between male (66 [95% CI 47–85]) and female (62 [46–78]) patients. There was a trend toward increased rate of visits among the black population with diabetes (92 [60–124]) compared with the white population with diabetes (59 [47–72]). Comparing the first 6 years of visits (1993–1998) with the last 5 years (1999–2003), we noted an increasing number of visits for DKA (Table 1) with a stable rate per 10,000 U.S. population with diabetes, from 59 (95% CI 42–76) to 69 (51–86).

We report the first national emergency department–based epidemiological study of DKA visits. Between 1993 and 2003, we found an increasing number of visits over time, but the rate remained relatively unchanged, suggesting that increased prevalence of diabetes accounts for the growth in DKA visits. The more recent estimate of 69 emergency department visits per 10,000 U.S. population with diabetes is lower than the CDC-reported rate of 84 hospital discharges for DKA during the same time period (3). Additionally, CDC data suggest a decrease in the rate of hospital discharges for DKA from 106 to 84 per 10,000 U.S. population with diabetes. These trends suggest an increased utilization of the emergency department in DKA management, a decreased incidence of DKA during hospital admission, or fewer “direct admissions” for DKA (bypassing the emergency department). The relative importance of these explanations requires further study.

Nevertheless, given its role at the front line of the acute health care system, the emergency department provides a unique perspective on DKA epidemiology. For example, we found that most emergency department visits for DKA led to hospital admission, with one in four to an ICU setting. This suggests that many DKA visits are sufficiently controlled during the emergency department visit to obviate a critical care admission, but clearly, emergency department visits for DKA require significant resource utilization. Future studies might focus on differences in clinical outcomes in the different disposition groups and cost analyses.

We found no significant difference in the rate of DKA per 10,000 emergency department visits by sex, race, ethnicity, or insurance status. We did, however, note a trend toward increased DKA visits among the overall and population with diabetes among blacks. The trend may be due, in part, to racial disparity or ketosis-prone type 2 diabetes in black patients (10). This finding will require continued surveillance.

Because the NHAMCS lacks identifiers, the data pertain to emergency department visits, not individual patients. Also, the NHAMCS data only contain information about actions during the emergency department visit, exclusive of prior events. Low sample size for a few subgroups precluded some comparisons. However, this combined database provides important information on outcomes that are known to be collected with accuracy.

In summary, we report the first national emergency department–based epidemiological study of DKA visits. The study provides excellent background data for future studies on resource utilization and for emergency department–based clinical trials. The emergency department provides a venue that offers a distinct and important perspective on diabetes and diabetes-related complications.

Table 1—

Emergency department visits with DKA, according to patient and hospital characteristics, 1993–2003

VariablenEstimated total no. of cases%95% CIRate per 10,000 U.S. population (95% CI)Rate per 10,000 ED visits (95% CI)
Year       
    1993–1998 88 315,000 42 32–52 2 (1–3) 6 (4–7) 
    1999–2003 122 438,000 58 48–68 3 (2–4) 8 (6–10) 
Age (years)       
    <10 8 NC NC NC NC NC 
    10–19 37 165,000 22 14–30 4 (2–5) 12 (7–17) 
    20–29 38 135,000 18 12–25 3 (2–5) 7 (4–10) 
    30–39 49 180,000 24 16–32 4 (2–5) 10 (7–14) 
    40–49 32 93,000 12 7–18 2 (1–3) 7 (4–10) 
    50–59 17 NC NC NC NC NC 
    60–69 10 NC NC NC NC NC 
    70–79 13 NC NC NC NC NC 
    ≥80 6 NC NC NC NC NC 
Sex       
    Male 96 370,000 49 39–59 3 (2–3) 7 (5–8) 
    Female 114 383,000 51 41–61 2 (2–3) 7 (5–9) 
Race       
    White 144 543,000 72 65–79 2 (2–3) 6 (5–8) 
    Black 61 180,000 24 17–31 5 (3–6) 8 (5–10) 
    Other 5 NC NC NC NC NC 
Ethnicity       
    Hispanic 18 NC NC NC NC NC 
    Non-Hispanic 154 551,000 73 65–81 2 (2–3) 8 (6–10) 
    Unknown 38 158,000 21 13–29 NA 7 (4–11) 
Insurance       
    Private 58 215,000 29 20–37 NA 5 (3–7) 
    Public 79 287,000 38 30–47 NA 8 (6–10) 
    Self-pay 33 119,000 16 9–23 NA 7 (4–10) 
    Other/unknown 40 132,000 17 11–24 NA 8 (5–12) 
Region       
    Northeast 40 112,000 15 9–21 2 (1–3) 5 (3–7) 
    Midwest 57 253,000 34 24–43 4 (2–5) 9 (6–12) 
    South 72 239,000 32 23–41 2 (1–3) 6 (4–8) 
    West 41 149,000 20 13–27 2 (1–3) 7 (5–10) 
Urban status       
    Urban 179 618,000 82 75–90 3 (2–3) 7 (6–9) 
    Nonurban 31 135,000 18 11–26 2 (1–3) 6 (3–8) 
Disposition       
    Admitted 174 651,000 87 81–92 NA 5 (4–6) 
    ICU 51 213,000 28 20–36 NA 14 (9–19) 
    Non-ICU 123 438,000 58 51–66 NA 5 (4–6) 
    Transferred 12 NC NC NC NA NC 
    Discharged 24 NC NC NC NA NC 
Total 210 753,000 100 — 3 (2–3) 7 (6–8) 
VariablenEstimated total no. of cases%95% CIRate per 10,000 U.S. population (95% CI)Rate per 10,000 ED visits (95% CI)
Year       
    1993–1998 88 315,000 42 32–52 2 (1–3) 6 (4–7) 
    1999–2003 122 438,000 58 48–68 3 (2–4) 8 (6–10) 
Age (years)       
    <10 8 NC NC NC NC NC 
    10–19 37 165,000 22 14–30 4 (2–5) 12 (7–17) 
    20–29 38 135,000 18 12–25 3 (2–5) 7 (4–10) 
    30–39 49 180,000 24 16–32 4 (2–5) 10 (7–14) 
    40–49 32 93,000 12 7–18 2 (1–3) 7 (4–10) 
    50–59 17 NC NC NC NC NC 
    60–69 10 NC NC NC NC NC 
    70–79 13 NC NC NC NC NC 
    ≥80 6 NC NC NC NC NC 
Sex       
    Male 96 370,000 49 39–59 3 (2–3) 7 (5–8) 
    Female 114 383,000 51 41–61 2 (2–3) 7 (5–9) 
Race       
    White 144 543,000 72 65–79 2 (2–3) 6 (5–8) 
    Black 61 180,000 24 17–31 5 (3–6) 8 (5–10) 
    Other 5 NC NC NC NC NC 
Ethnicity       
    Hispanic 18 NC NC NC NC NC 
    Non-Hispanic 154 551,000 73 65–81 2 (2–3) 8 (6–10) 
    Unknown 38 158,000 21 13–29 NA 7 (4–11) 
Insurance       
    Private 58 215,000 29 20–37 NA 5 (3–7) 
    Public 79 287,000 38 30–47 NA 8 (6–10) 
    Self-pay 33 119,000 16 9–23 NA 7 (4–10) 
    Other/unknown 40 132,000 17 11–24 NA 8 (5–12) 
Region       
    Northeast 40 112,000 15 9–21 2 (1–3) 5 (3–7) 
    Midwest 57 253,000 34 24–43 4 (2–5) 9 (6–12) 
    South 72 239,000 32 23–41 2 (1–3) 6 (4–8) 
    West 41 149,000 20 13–27 2 (1–3) 7 (5–10) 
Urban status       
    Urban 179 618,000 82 75–90 3 (2–3) 7 (6–9) 
    Nonurban 31 135,000 18 11–26 2 (1–3) 6 (3–8) 
Disposition       
    Admitted 174 651,000 87 81–92 NA 5 (4–6) 
    ICU 51 213,000 28 20–36 NA 14 (9–19) 
    Non-ICU 123 438,000 58 51–66 NA 5 (4–6) 
    Transferred 12 NC NC NC NA NC 
    Discharged 24 NC NC NC NA NC 
Total 210 753,000 100 — 3 (2–3) 7 (6–8) 

ED, emergency department; NA, not available; NC, not calculable due to n < 30 (see research design and methods).

View Large
1.
Javor KA, Kotsanos JG, McDonald RC, Baron AD, Kesterson JG, Tierney WM: Diabetic ketoacidosis charges relative to medical charges of adult patients with type 1 diabetes.
Diabetes Care
 
20
:
349
–354,
1997
2.
Basu A, Close CF, Jenkins D, Krentz AJ, Nattrass M, Wright AD: Persisting mortality in diabetic ketoacidosis.
Diabet Med
 
10
:
282
–284,
1993
3.
National Diabetes Surveillance System: Diabetic ketoacidosis as first-listed diagnosis [article online],
2005
. Available from http://www.cdc.gov/diabetes/statistics/dkafirst/fig1.htm. Accessed 18 March
2006
4.
Freire AX, Umpierrez GE, Afessa B, Latif KA, Bridges L, Kitabchi AE: Predictors of intensive care unit and hospital length of stay.
J Crit Care
 
17
:
207
–211,
2002
5.
Marinac JS, Mesa L: Using a severity of illness scoring system to assess intensive care unit admissions for diabetic ketoacidosis.
Crit Care Med
 
28
:
2238
–2241,
2000
6.
McCaig LF, Burt CW: National Hospital Ambulatory Medical Care Survey: 2003 emergency department summary.
Adv Data
 
358
:
1
–37,
2005
7.
Stussman BJ: National Hospital Ambulatory Medical Care Survey: 1993 emergency department summary.
Adv Data
 
271
:
1
–15,
1996
8.
McCaig LF, McLemore T: Plan and operation of the National Hospital Ambulatory Medical Survey. Series 1: Programs and collection procedures.
Vital Health Stat
 
34
:
1
–78,
1994
9.
Geiss LS, Herman WH, Goldschmid MG, DeStefano F, Eberhardt MS, Ford ES, German RR, Newman JM, Olson DR, Sepe SJ, Stevenson JM, Vinicor F, Wetterhall SF, Will JC: Surveillance for diabetes mellitus: United States, 1980–1989.
MMWR CDC Surveill Summ
 
42
:
1
–20,
1993
10.
Umpierrez GE, Smiley D, Kitabchi AE: Narrative review: ketosis-prone type 2 diabetes mellitus.
Ann Intern Med
 
144
:
350
–357,
2006

A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

DIABETES CARE
2006