Methicillin-resistant Staphylococcus aureus (MRSA) is
increasingly a cause of nosocomial and community-onset
infection with unknown national scope and magnitude. We
used the National Hospital Discharge Survey to calculate
the number of US hospital discharges listing S. aureus–spe-
cific diagnoses, defined as those having at least 1
International Classification of Diseases (ICD)-9 code specif-
ic for S. aureus infection. The number of hospital discharges
listing S. aureus–specific diagnoses was multiplied by the
proportion of methicillin resistance for each corresponding
infection site to determine the number of MRSA infections.
From 1999 to 2000, an estimated 125,969 hospitalizations
with a diagnosis of MRSA infection occurred annually,
including 31,440 for septicemia, 29,823 for pneumonia, and
64,706 for other infections, accounting for 3.95 per 1,000
hospital discharges. The method used in our analysis may
provide a simple way to assess trends of the magnitude of
MRSA infection nationally.
most common causes of healthcare-associated infections
reported to the National Nosocomial Infections
Surveillance (NNIS) System, including ventilator-associ-
ated pneumonia, surgical site infection, and catheter-asso-
ciated bloodstream infection (1). S. aureus is also a
frequent cause of community-associated infections, partic-
ularly skin and soft tissue infections. Although most com-
munity-onset infections are treated in the outpatient
setting, some invasive infections, including bacteremia,
septic arthritis, toxic shock syndrome, osteomyelitis, and
endocarditis, have devastating complications and may
require hospitalization (2).
taphylococcus aureus is a major cause of infection in
both healthcare and community settings. It is one of the
Antimicrobial resistance in S. aureus emerged soon
after penicillin came into common use in the 1940s.
During the next 2 decades, resistance of this pathogen to
penicillin became widespread, followed by increasing
resistance to the new semisynthetic penicillinase-resistant
antimicrobial drugs (e.g., methicillin, oxacillin, nafcillin)
(3). In the last 20 years, methicillin-resistant S. aureus
(MRSA) has spread throughout the world in healthcare set-
tings, leading to an increased reliance on vancomycin for
empiric treatment (4). Recently, S. aureus resistance to
vancomycin, the last commonly used antimicrobial drug to
which this organism was considered uniformly suscepti-
ble, has emerged (5). In addition, serious MRSA infection
has been increasingly reported in persons without identi-
fied predisposing risk, including recent healthcare expo-
MRSA infections are thought to cause substantial ill-
ness and contribute to healthcare costs in the United States.
However, published estimates vary widely and have been
based on single-center or local data with limited applica-
bility (4,7). Accurate estimates of the incidence of MRSA
infection are essential to determine effects on health and
healthcare expenditures. Since most patients with serious
MRSA infections are hospitalized, we focused our esti-
mate on hospitalized patients.
The incidence of S. aureus infection was estimated
from the number of hospitalizations with S. aureus–related
discharge diagnoses in a national surveillance database.
We used 1999 and 2000 public-use data from the National
Hospital Discharge Survey (NHDS) to calculate the num-
ber of hospital discharges with at least 1 S. aureus–related
discharge diagnosis. All acute-care hospitalizations, except
infants whose hospital stay began at their own birth, were
S taphyloc oc c us aureus
Hospitalizations, United S tates
Matthew J. Kuehnert,* Holly A. Hill,* Benjamin A. Kupronis,* Jerome I. Tokars,* Steven L. Solomon,*
and Daniel B. Jernigan*
868 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 11, No. 6, June 2005
*Centers for Disease Control and Prevention, Atlanta, Georgia,
included. The NHDS is a nationally representative annual
sample of discharge records from ≈475 nonfederal short-
stay hospitals (8). The survey is based on a stratified, mul-
tistage probability design; the sampled hospital discharge
records are weighted to produce national estimates. The
database includes ≤7 principal discharge diagnoses. We
identified S. aureus–related discharge diagnoses by using
the International Classification of Diseases, Ninth
Revision, Clinical Modification (ICD-9-CM) codes specif-
ic for S. aureus infection: 038.11 (S. aureus septicemia),
482.41 (S. aureus pneumonia), and 041.11 (S. aureus
infection in conditions classified elsewhere or of unspeci-
fied site). A discharge record listing multiple S. aureus-
related diagnoses was counted only once. Septicemia was
preferentially included, followed by S. aureus–related
Next, the percentage of isolates resistant to oxacillin
was determined. To simplify terminology, resistance to
methicillin and oxacillin hereafter are used interchange-
ably. Oxacillin is used as a proxy for testing of susceptibil-
ity to all β-lactam antimicrobials, including methicillin.
The Surveillance Network (TSN) Database-USA (Focus
Technologies, Herndon, VA, USA) was the source of
antimicrobial susceptibility testing results. TSN is a repos-
itory of quantitative and qualitative susceptibility results
collected from >200 microbiology laboratories in the
United States. These laboratories make up a nationally rep-
resentative sample based on associated hospital bed size,
patient population, and geographic region as determined by
the US Bureau of the Census (9,10). Susceptibility testing
of patient isolates is conducted on site by each participating
laboratory as part of routine diagnostic testing; only isolates
judged as clinically significant are included. Data are gen-
erated by using Food and Drug Administration–approved
testing methods. S. aureus antimicrobial susceptibility to
oxacillin was classified as susceptible, intermediate, or
resistant according to NCCLS breakpoint criteria; we clas-
sified intermediate isolates as methicillin-susceptible S.
aureus for purposes of this analysis. Data were stratified by
site of infection, i.e., bloodstream, lung, and other sites.
Duplicate isolates were removed if the initial and subse-
quent isolates were cultured within 30 days of each other.
The number of hospital discharges listing S.
aureus–specific diagnoses was multiplied by the propor-
tion of methicillin resistance at each corresponding infec-
tion site to determine the total number of MRSA
infections. Infections also were stratified by geographic
region and age. The frequency of primary diagnosis and
the 10 most frequent secondary (all-listed) diagnoses were
abstracted from hospitalizations that included S.
Results for the years 1999–2000 were determined by
calculating data specific to each year and then averaging.
Data on resistance rates were stratified first by region and
then by age; for each stratification, a chi-square test was
used to determine whether differences were significant.
The Cochran-Armitage test, a nonparametric method, was
used to determine the trend in MRSA hospitalization rate
by age category.
The effects of region and age on the incidence rate of
MRSAwere assessed by calculating relative rates and their
associated 95% confidence intervals, with the lowest rates
designated as comparison groups. Since the rate of S.
aureus and the MRSAproportion were estimated separate-
ly and then multiplied to obtain the MRSA hospitalization
rate, the variance of the MRSA rate was calculated by
using the delta method (11). The variance of the methi-
cillin resistance proportions was determined under the
assumption that the antimicrobial susceptibility data
reflected those that would have been derived from a ran-
dom sample of all S. aureus isolates in the United States in
that time period. Variance estimates were calculated using
SUDAAN software (Research Triangle Institute, Research
Triangle Park, NC, USA). For both S. aureus rates and
methicillin resistance proportions, variances were estimat-
ed separately for 1999 and 2000, and the larger of the vari-
ance estimates was used in subsequent calculation of 95%
confidence intervals for relative rates.
We estimate that 291,542 hospital discharges with S.
aureus infection-related diagnoses occurred annually from
1999 to 2000 (Table 1). A diagnosis of S. aureus infection
occurred in 9.13 of every 1,000 hospital discharges. The
overall rate of methicillin resistance for all S. aureus infec-
tions was reported to be 43.2%. MRSA rates for sep-
ticemia, pneumonia, and other infections increased with
patient age. An estimated 125,969 hospitalizations with 1
or more discharge diagnoses associated with MRSAinfec-
tion occurred annually, accounting for 3.95 of every 1,000
hospital discharges. For all sites, most diagnoses occurred
in persons ≥65 years of age.
In hospitalizations in which S. aureus septicemia and
pneumonia were listed as discharge diagnoses, these con-
ditions were primary diagnoses in 34.3% and 49.3% of dis-
charges, respectively. For S. aureus infection in conditions
classified elsewhere and in an unspecified site, a diagnosis
intended only for secondary listing, the most frequent pri-
mary diagnoses were postoperative (e.g., wound) infection
(10.1%), cellulitis or abscess (9.9%), infection from an
implanted device or graft (7.3%), and urinary tract infec-
The largest proportion of S. aureus–related discharge
diagnoses occurred in patients from the South, followed by
the Midwest, Northeast, and West (Table 2). For both, the
rate of S. aureus discharge diagnoses and methicillin
MRSA Hospitalizations, United States
Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 11, No. 6, June 2005869
resistance proportion, significant differences were seen by
geographic region. S. aureus discharge diagnoses were sig-
nificantly higher for the South than the Northeast, while
for methicillin resistance proportion, the Northeast,
Midwest, and South were significantly higher than the
West (p<0.05 for all comparisons). The South had the
highest MRSA hospitalization rate, reflecting both the S.
aureus rate and methicillin resistance proportion, which
was significantly higher than the MRSA rate estimated for
the West (South vs. West, relative risk 1.57, 95% confi-
dence interval 1.29–1.91).
Most S. aureus–related discharge diagnoses occurred in
patients ≥65 years of age. When S. aureus diagnoses by
rate were examined, a bimodal distribution was seen, with
highest rates occurring in children and the elderly (Table
3). Patients ≤14 and 15–44 years of age had higher MRSA
hospitalization rates compared with patients 45–64 and
≥65 years of age (p<0.01). Overall, the MRSA rate
increased with patient age (p<0.05 for trend).
Infectious diseases cause many hospitalizations each
year in the United States; these diseases include syndromes
commonly associated with S. aureus. In 1994, the rate of
hospitalization for infectious disease was 15 per 1,000 US
population, with a total of 4 million hospitalizations,
including 1,480,000 pneumonias, 335,000 skin infections,
and 302,000 septicemias; yearly rates for these disease
syndromes were similar from 1999 to 2000 (12–14).
Gram-positive organisms are an increasingly recognized
cause of systemic infection, including sepsis (12,15). More
than half of all sepsis cases are estimated to be caused by
gram-positive organisms, including S. aureus (16). In the
Calgary Health Region in Canada, the annual incidence of
invasive S. aureus infection was estimated to be 28.4 cases
per 100,000 population from 1999 to 2000, which is com-
parable with the rate of invasive pneumococcal disease and
exceeds the rate of invasive streptococcal infection (17).
Drug resistance in S. aureus, including the emergence
of MRSA in healthcare and community settings, is an
increasingly reported event that makes treating serious
infection difficult. Extrapolating from our estimates and
those of Simonsen et al. (12), a rate of ≈47 diagnoses per
100,000 population, making up 3% of all infectious dis-
ease hospitalizations, were associated with laboratory-con-
firmed MRSA infection from 1999 to 2000, and ≈10% of
septicemias were caused by MRSA.
Although the burden of MRSA infection has not been
systematically estimated nationally, past estimates have
been based on single-center or selected population-based
studies in the United States. Based on ICD-9-CM data
from the New York City metropolitan area, an estimated
1.0% of hospital discharges are associated with S. aureus
infection, and 0.21% of discharges are estimated to be
associated with MRSA (18). In 1995, based on extrapola-
tion of hospital discharge data from NHDS and nosocomi-
al infection data from the NNIS System, an estimated
206,504 S. aureus infections (0.58% of admissions) and
70,270 MRSA infections (0.20% of admissions) were
acquired in the healthcare setting (Centers for Disease
870Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 11, No. 6, June 2005
Control and Prevention, unpub. data). Our estimates for
1999 to 2000 are similar for S. aureus infections but are
higher for MRSA.
Although ICD-9-CM coding accuracy for S. aureus
infections has not been specifically examined, the accura-
cy of coding for sepsis from all causes has been reviewed,
and has demonstrated a sensitivity >75% for any sep-
ticemia or bacteremia code and positive and negative pre-
dictive values >80% for the code specific for
Staphylococcus spp. septicemia (ICD-9-CM 038) (16,19).
However, the relationship between true S. aureus infec-
tions and ICD-9 discharge coding should be further
assessed to validate this method as a tool for monitoring
We found associations between MRSA rate and both
region and age. This finding is consistent with previously
published data showing an association between age and
both the incidence of invasive S. aureus infection and the
rate of methicillin resistance (17,20). We also demonstrat-
ed a significant difference in MRSA discharge rates
between the South and West. Although past microbiologic
surveys also have reported higher rates of methicillin
resistance in the South compared with other regions, the
reasons for this variation are unclear (21,22). These differ-
ences may need to be assessed as community-associated
MRSA infection becomes more common.
Our estimate is subject to a number of limitations that
most likely underestimated hospitalizations associated
with MRSA infection. First, S. aureus infections may not
have been accurately represented by the ICD-9-CM dis-
charge code; colonization may have been inadvertently
included; and more likely, true infections may not have
been identified, since these diagnoses require laboratory
culture confirmation. Since only 7 principal diagnoses are
included in NHDS, infections listed less prominently may
have been excluded. Duplicate isolates were excluded
when identified within 30 days of each other; thus, unusu-
al scenarios, such as multiple infections during a hospital-
ization or infections present for >30 days, were not
included. We were not able to distinguish between commu-
nity- and healthcare-acquired infection. However, this
analysis was designed to measure the overall incidence of
disease associated with acute care hospitalization, regard-
less of acquisition site, and did not include disease man-
aged in the outpatient setting. Although previously pub-
lished region and age stratification groups were used,
which reduces risk of bias, unmeasured confounders may
have affected calculated trends. Finally, although both
NHDS and TSN data aim to represent nationally represen-
tative samples based on similar factors, methods may have
differed, which could have skewed our results. For all data
used, institutional settings, such as long-term care or cor-
rectional facilities, were not included.
In summary, our estimates indicate that the national
burden of serious MRSA disease is quantifiable and sub-
stantial. Measurement of trends in S. aureus disease, such
as the increasing incidence of antimicrobial resistance
associated with certain age groups and geographic regions,
will have implications in the development of prevention
programs, both in the healthcare and community settings.
Our method provides a simple way to estimate trends of
magnitude of hospitalization associated with S. aureus
infection in the United States and could complement meth-
ods currently in place for national surveillance.
We appreciate invaluable technical assistance provided by
Maria Owings, Jean Kozak, Ron Master, and Daniel Sahm.
Dr. Kuehnert is a medical epidemiologist at the National
Center for Infectious Diseases, Centers for Disease Control and
Prevention. His research interests have included antimicrobial
resistance surveillance and now focus on improvement of blood,
organ, and other tissue safety.
1. National Nosocomial Infections Surveillance (NNIS) system report,
data summary from January 1992–June 2001, issued June 2001. Am
J Infect Control. 2001;29;404–21.
2. Lowy FD. Staphylococcus aureus infections. N Engl J Med.
3. Finland M. Emergence of antibiotic resistance in hospitals. Rev Infect
4. Jernigan JA, Clemence MA, Stott GA, Titus MG, Alexander CH,
Palumbo CM, et al. Control of methicillin-resistant Staphylococcus
aureus at a university hospital: one decade later. Infect Control Hosp
5. Chang S, Sievert DM, Hageman JC, Boulton ML, Tenover FC,
Downes FP, et al. Infection with vancomycin-resistant
Staphylococcus aureus containing the vanA resistance gene. N Engl J
MRSA Hospitalizations, United States
Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 11, No. 6, June 2005 871
6. Herold BC, Immergluck LC, Maranan MC, Lauderdale DS, Gaskin
RE, Boyle-Vavra S, et al. Community-acquired methicillin-resistant
Staphylococcus aureus in children with no identified predisposing
risk. JAMA. 1998;279:593–8.
7. Petti CA, Fowler VG Jr. Staphylococcus aureus bacteremia and
endocarditis. Infect Dis Clin North Am. 2002;16:413–35.
8. Dennison C, Pokras R. Design and operation of the National Hospital
Discharge Survey: 1988 redesign. Vital Health Stat. 2000;1:1–42.
9. Jones ME, Mayfield DC, Thornsberry C, Karlowsky JA, Sahm DF,
Peterson D. Prevalence of oxacillin resistance in Staphylococcus
aureus among inpatients and outpatients in the United States during
2000. Antimicrob Agents Chemother. 2002;46:3104–5.
10. U.S. Bureau of the Census. Census 2000 geographic terms and con-
cepts. Washington; 2000.
11. Oehlert GW. A note on the delta method. Am Stat. 1992;46:27–9.
12. Simonsen L, Conn LA, Pinner RW, Teutsch S. Trends in infectious
disease hospitalizations in the United States, 1980–1994. Arch Intern
13. Popovic JR, Hall MJ. 1999 National Hospital Discharge Survey.
Advance data from vital and health statistics. No. 319. Hyattsville
(MD): National Center for Health Statistics; 2001.
14. Hall MJ, Owings MF. 2000 National Hospital Discharge Survey.
Advance data from vital and health statistics. No. 329. Hyattsville
(MD): National Center for Health Statistics; 2002.
15. Bone RC. Gram-positive organisms and sepsis. Arch Intern Med.
16. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of
sepsis in the United States from 1979 through 2000. N Engl J Med.
17. Laupland KB, Church DL, Mucenski M, Sutherland LR, Davies HD.
Population-based study of the epidemiology of and the risk factors for
invasive Staphylococcus aureus
18. Rubin RJ, Harrington CA, Poon A, Dietrich K, Greene JA,
Moiduddin A. The economic impact of Staphylococcus aureus infec-
tion in New York City hospitals. Emerg Infect Dis. 1999;5:9–17.
19. Ollendorf DA, Fendrick AM, Massey K, Williams GR, Oster G. Is
sepsis accurately coded on hospital bills? Value Health.
20. Diekema DJ, Pfaller MA, Jones RN, and the SENTRY Participants
Group. Age-related trends in pathogen frequency and antimicrobial
susceptibility of bloodstream isolates in North America. Int J
Antimicrob Agents. 2002;20:412–8.
21. Wakefield DS, Pfaller M, Massanari RM, Hammons GT. Variation in
methicillin-resistant Staphylococcus aureus occurrence by geograph-
ic location and hospital characteristics. Infect Control. 1987;8:151–7.
22. 1997 ASCP Susceptibility Testing Group. United States geographic
bacteria susceptibility patterns. Diag Microbiol Infect Dis.
infections. J Infect Dis.
Address for correspondence: Daniel B. Jernigan, Division of Healthcare
Quality Promotion, National Center for Infectious Diseases, Centers for
Disease Control and Prevention, 1600 Clifton Rd NE, Mailstop A35,
Atlanta, GA30333, USA; fax: 404-639-2647; email: email@example.com
872Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 11, No. 6, June 2005