This study was based on data from the Healthcare Cost and Utilization Project National Inpatient Sample (HCUP NIS) for the 5-year period 2000–2004. HCUP NIS is a stratified random sample from ≈20% of all US community hospitals. For 2004, the NIS contains all discharge data from >1,000 acute-care hospitals in 37 states, representing >8 million hospital stays.

We identified all hospital admissions in the HCUP NIS for which a principal diagnosis of SSTI was given. As defined by the Uniform Hospital Discharge Data Set, principal diagnosis is the condition “established after study to be chiefly responsible for occasioning the admission of the patient to the hospital for care” (http://wonder.cdc.gov).

SSTIs were defined as the following: 1) acute lymphadenitis (International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM] 683.XX); 2) carbuncle and furuncle (680.XX); 3) cellulitis and abscess of finger and toe (681.XX); 4) impetigo (684.XX); 5) infection (chronic) of amputation stump (997.62); 6) other cellulitis and abscess (682.XX); 7) other local infections of skin and subcutaneous tissue (686.XX); 8) abscess of anal and rectal regions (566); 9) chronic ulcer of other specified sites (707.8); 10) chronic ulcer of unspecified site (707.9); 11) decubitus ulcer (707); 12) infection due to other vascular device implant and graft (996.62); 13) pilonidal cyst with abscess (685); 14) postoperative wound infection (998.5X); 15) posttraumatic wound infection, not elsewhere classified (958.3); 16) ulcers of lower limbs, except decubitus (707.1X); 17) gangrene (785.4); and 18) necrotizing fasciitis (728.86).

To aid in interpretation of the data, we grouped the above-listed SSTIs into 3 mutually exclusive categories: 1) superficial infections predominantly caused by S. aureus or Streptococcus pyogenes (groups 1–7 above); 2); deeper or healthcare-associated infections more likely to involve anaerobic or gram-negative organisms (groups 8–16); and 3) infections typically associated with a high rate of mortality (groups 17–18). For each year of interest, all SSTI admissions were then stratified by type of infection and selected patient and hospital characteristics.

To provide a benchmark against which to interpret possible trends in SSTI-related hospital admissions, we generated similar series for all admissions with a principal diagnosis of infectious pneumonia, another common type of infection that frequently results in hospitalization but for which no trends were expected. Infectious pneumonia was defined to include 1) pneumococcal pneumonia (ICD-9-CM 481); 2) other bacterial pneumonia (482.XX) 3); pneumonia caused by other specified organism (483.X) 4); bronchopneumonia, organism unspecified (485); and 5) pneumonia, organism unspecified (486).

Analyses were directed at discerning possible trends in SSTI-related hospital admissions at US acute-care hospitals over the 5-year period of study. Total SSTI admissions were estimated for each year by type of infection (i.e., superficial, deeper or healthcare-associated, often-fatal). Results also were stratified according to patient (e.g., age, sex) and hospital characteristics (e.g., urban vs. rural) available in the HCUP NIS. For comparison, we also examined total hospital admissions for infectious pneumonia.

Because the HCUP NIS undergoes periodic changes in its sampling frame and weighting method, as well as in data elements and definitions, we used the NIS Trends Supplemental (NIS-Trends) files to ensure comparability of data across the years of interest. The NIS-Trends files contain sampling weights and data elements that are consistently defined across all years of interest.

Analyses were primarily descriptive (i.e., no a priori hypotheses were made, and our results were not subjected to formal significance testing). All analyses were conducted by using the SAS Proprietary Software, Release 9.1 (SAS Institute, Cary, NC, USA).

The estimated total number of annual SSTI admissions to US acute-care hospitals rose steadily over the 5-year period, from 675,000 in 2000 to 869,800 in 2004, an increase of 194,000 (29%) admissions. In contrast, total admissions for infectious pneumonia fluctuated from year to year and were largely unchanged over this period (Table 1).

The increase in SSTI admissions was greatest among younger (age <65 years) rather than older patients (age 65–100 years) (37% vs. 14%, respectively) and for urban rather than rural hospitals (32% vs. 11%) (Table 2). The increase in SSTI admissions also was greatest among patients with superficial infections (e.g., cellulitis, abscess) rather than deeper or healthcare-associated infections (e.g., postoperative wound infection, infection due to vascular device) (33% vs. 24%).

Until recently, CA-MRSA was relatively uncommon. In 2000, for example, these isolates accounted for only 3% of staphylococcal isolates submitted to Minnesota laboratories (4). In recent years, however, CA-MRSA has become a major cause of SSTIs. The prevalence of MRSA among patients with SSTIs who sought treatment at 1 Los Angeles area emergency department increased from 29% in 2001 to 64% in 2004 (5).

While the estimated total number of SSTI admissions to US acute-care hospitals increased by ≈29% during 2000–2004, admissions for infectious pneumonia were largely unchanged. We are unaware of any other data to which our findings might be directly compared. Our results appear to be consistent, however, with those of previous studies that have reported an increasing prevalence of illness attributable to S. aureus infections in general and to MRSA in particular (6,7). Although we could not establish in our study whether the increase in the number of SSTI hospital admissions was a result of the growing prevalence of CA-MRSA (the HCUP-NIS does not report microbiologic data), we suspect that the 2 phenomena are closely linked—especially in light of the absence of any similar increase in hospital admissions for pneumonia, the most common community-associated infection requiring hospitalization. We therefore believe that the clinical and economic effects of CA-MRSA SSTIs are substantial and growing, and that this increase should be a focus of additional research.