Share this article on:

HIV Seropositivity Predicts Longer Duration of Stay and Rehospitalization Among Nonbacteremic Febrile Injection Drug Users With Skin and Soft Tissue Infections

Hsieh, Yu-Hsiang PhD, MSc*; Rothman, Richard E MD, PhD*†; Bartlett, John G MD; Yang, Samual MD*; Kelen, Gabor D MD*

JAIDS Journal of Acquired Immune Deficiency Syndromes: December 1st, 2008 - Volume 49 - Issue 4 - p 398-405
doi: 10.1097/QAI.0b013e318183ac84
Clinical Science

Background: Skin/soft tissue infections (SSTIs) are the leading cause of hospital admissions among injection drug users (IDUs).

Methods: We performed a retrospective investigation to determine the epidemiology of SSTIs (ie, cellulitis and/or abscesses) in febrile IDUs, with a focus on bacteriology and potential predictors of increased health care utilization measured by longer length of stay and rehospitalization. Subjects were drawn from a cohort of febrile IDUs presenting to an inner-city emergency department from 1998 to 2004.

Results: Of the 295 febrile IDUs with SSTIs, specific discharge diagnoses were cellulitis only (n = 143, 48.5%), abscesses only (n = 113, 38.3%), and both (n = 39, 13.2%). Documented HIV infection rate was 28%. Of note, 10 subjects were newly diagnosed with HIV infection during their visits. Staphylococcus aureus was the leading pathogen, and increasing rates of methicillin-resistant S. aureus emerged over time (before 2001: 4%, 2001-2004: 56%, P < 0.01). HIV seropositivity predicted rehospitalization within 90 days [adjusted hazard ratios and 95% confidence intervals: 2.90 (1.20 to 7.02)]. HIV seropositivity also predicted increased length of stay in those who were nonbacteremic [adjusted hazard ratios and 95% confidence intervals: 1.49 (1.11 to 2.01)].

Conclusions: Among febrile IDUs with SSTIs, a strong association between HIV seropositivity and health care resource utilization was found. Accordingly, attention to HIV serostatus should be considered in clinical disposition decisions for this vulnerable high-risk population.

From the *Department of Emergency Medicine; and †Division of Infectious Diseases, Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, MD.

Received for publication February 4, 2008; accepted May 16, 2008.

Supported in part by Grants (R03-HS-009555 and M01-RR-000052) from the National Institutes of Health (R.E.R.).

Presented in part at the Society for Academic Emergency Medicine Annual Meeting, May 2006, San Francisco, CA.

Potential conflicts of interest: J.G.B. has served on HIV Advisory Boards for Abbott, Glaxo, Bristol, Pfizer, and Tibotec and Policy Board for Johnson and Johnson. All other authors: no conflicts.

Correspondence to: Dr. Yu-Hsiang Hsieh, PhD, MSc, Department of Emergency Medicine, The Johns Hopkins University, 5801 Smith Avenue, Suite 3220, Davis Building, Baltimore, MD 21209 (e-mail:

Back to Top | Article Outline


There are an estimated 13 million injection drug users (IDUs) worldwide,1 making it an important social and medical problem. Injection drug use (IDU) increases the risk of acquiring a variety of infectious diseases ranging from sexually transmitted infections to bacteremia.2 The most common IDU-related infectious disease complications are skin and soft tissue infections (SSTIs), which include cellulitis and abscesses3,4 with reported prevalence of 22%-32%.5,6 In a cohort study from Amsterdam, the incidence of abscesses among IDUs was reported as high as 33 per 100 person-years.7 In the United States, SSTIs are the leading cause (17%-18%) for emergency department (ED) visits or revisits in IDUs and the second leading cause (18%) of hospitalizations.4,8 A higher prevalence of abscesses and other SSTIs has been linked with the use of black tar heroin9 which is predominant in the West Coast of the United States but not in the northeast or mid-Atlantic regions.10,11 Hospitalization due to IDU-related soft tissue infections has been associated with some sociodemographic and health care utilization factors, such as residing in urban metropolitan area,12 living in shelters, and frequent hospitalization in the past year.13 The health care (ED and inpatient) costs associated with SSTIs in this high-risk population are substantial, with 1 study estimating over 20 million dollars per year spent at a single hospital.14

Historically, the most common bacterial pathogens causing SSTIs in IDUs are Staphylococcus aureus and streptococci (group A, C, F, and G).15-18 Other bacteria including Enterococcus faecalis, viridans group streptococci, coagulase-negative staphylococci, anaerobes, Enterobacteriaceae, and Eikenella corrodens are present but less common. Most recently, rapid changes in the bacteriology of community-acquired SSTIs have been reported in the general population, driven mainly by the emergence of methicillin-resistant Staphylococcus aureus (MRSA).19-21 In IDUs, a high-risk group for SSTIs, MRSA has previously been linked to local outbreaks of SSTIs in Detroit22,23 and Boston24 (early 1980s) and Zurich, Switzerland (late 1990s).25

Although the risk factors associated with SSTIs in IDUs have been extensively described,3 few, if any studies, have examined predictors of increased health care resource utilization, specifically longer duration of hospital stay and rehospitalization. Furthermore, there are few recent studies describing the bacteriology of SSTIs in IDUs, particularly relevant in light of increasing rates of drug resistance pathogens and efforts to optimize use of health care resources. The primary objective of this study was to determine the epidemiology of SSTIs (ie, cellulitis and/or abscesses) in febrile IDUs, with a focus on bacteriology and potential predictors of increased health care utilization measured by longer length of stay (LOS) and rehospitalization within 90 days.

Back to Top | Article Outline


Settings and Study Population

The Johns Hopkins Hospital Emergency Department (JHH ED) resides in an academic tertiary care facility located in inner city, East Baltimore, serving 50,000-55,000 patients per year during the 6-year study period. The ED population provides care principally to the local socioeconomically disadvantaged population with 30% uninsured, 31% recipients of public health insurance, 10%-12% HIV infected, and 15% IDUs.

Back to Top | Article Outline

Study Design and Patient Enrollment

Secondary data analysis study was conducted using data from a 6-year prospective study evaluating predictors and the outcomes of serious infections (principally infective endocarditis) among febrile IDUs. The main study began in March 1998, with consecutive, 24 hours a day and 7 days a week, enrollment of IDUs presenting to the JHH ED with febrile illness.26 The study protocol was approved by The Johns Hopkins University School of Medicine's Institutional Review Board.

According to the practice policy in JHH ED at the time this study was being conducted, all febrile IDUs were admitted to the inpatient unit for further evaluation due to their risks of bacteremia and infective endocarditis. For study purposes, to ensure accurate collection of data, those patients who provided consent were admitted to a dedicated research unit (General Clinical Research Center), which also serves as one of the general inpatient care units for their evaluation. Inclusion criteria were 18 years and older, documented ED fever of ≥100.4°F, self-reported IDU in ≤3 months, and ability to provide informed consent. Exclusion criteria were as follows: admission to the intensive care unit, belonging to the hospital HIV specialty care service (which admits patients only to a distinct inpatient unit), discharge or transfer from the ED, or left ED against medical advice. Each patient received standard of care workup in the ED for febrile illnesses in IDUs, including a complete history and physical examination, complete blood count, chest radiograph, and 2 sets of blood cultures before initiation of antibiotics. Hospital protocol dictated that 1 additional set of blood cultures was performed upon admission, with additional sets drawn, based on physician discretion and dependent on clinical course of the individual patient. Skin, soft tissue (SST) (skin/abscess/fluid/wound) culture was performed according to the clinical discretion of the ED or inpatient providers. The final clinical diagnosis was defined as the discharge diagnosis by the treating inpatient physicians. If the final clinical diagnosis of an enrollee in the main study was cellulitis and/or abscesses, the subject was included for data analysis in this study.

Back to Top | Article Outline

Data Collection

Demographics (age, sex, and race), comorbid conditions [documented HIV serostatus, CD4 cell counts, HIV RNA viral load, chronic hepatitis C virus (HCV) and hepatitis B virus (HBV) infection, and diabetes], blood and SST culture results, length of hospital stay, rehospitalization within 90 days, and final clinical diagnosis of the visit were collected. Data were obtained from the electronic patient record (EPR) database. Only information of CD4 and HIV viral load during the admission or within 6 months before the admission was obtained. Length of hospital stay was calculated as the number of days that the patients stayed in the hospital after admission. Newly diagnosed HIV infection was defined as serologic evidence of HIV infection during the visit in patients who denied prior documented HIV testing or there was absence of a prior positive test in the EPR. Culture data were recorded as reported from the EPR.

Back to Top | Article Outline

Data Analysis

This secondary data analysis was limited to the first SSTIs record for each IDU in the database. Patients who left against medical advice were excluded from the predictors for increased LOS data analyses. Coagulase-negative staphylococci were discounted as contaminants if they were recovered in 1 blood culture.27

Time-to-event analyses were used to determine the predictors for increased LOS. Bivariate analyses were first performed, followed by multivariate Cox proportional hazards regression analyses to identify predictors associated with increased LOS.28 Effect-measure modification between HIV infection and other covariates on LOS was assessed by stratified regression analysis. Febrile or SSTI rehospitalization rate within 90 days after discharge from an SSTI was defined as the number of febrile or SSTI rehospitalizations within 90 days divided by the number of SSTIs in a particular group. Kaplan-Meier analysis was performed to estimate the time to rehospitalization, and the log-rank test was used to detect significant difference between subgroups. To identify predictors for febrile and/or SSTIs readmission within 90 days after discharge, χ2 tests were performed first for the bivariate analysis. Effects of potential predictors on rehospitalization rates were examined by the multivariate Cox proportional hazards regression model analysis with the consideration of time to rehospitalization.

Hazard ratios and their 95% confidence intervals were calculated, and P values of <0.05 were considered as significant. All statistical tests were performed by using SAS version 9.1 (SAS Institute, Inc, Cary, NC).

Back to Top | Article Outline


Overall, 1164 febrile IDUs were approached during the 6-year study period. Among these, 849 (72.9%) provided consent and were enrolled. Among 315 nonenrollees, 172 declined, 131 belonged to the HIV specialty service, 7 left against medical advice in ED, 4 were transferred to other institutions, and 3 were admitted to intensive care unit. There were no statistical differences in age, sex, and race between those enrolled and those who declined.

The leading clinical diagnosis among the 849 subjects was cellulitis and/or abscesses (n = 301, 35.5%), followed by unknown illness (16.9%), community-acquired pneumonia (13.5%), and infective endocarditis (10.5%). Of those with SSTIs, 295 (98.0%) had complete information, allowing for more detailed descriptive analyses.

Among the 295 subjects with SSTIs, the majority were African Americans (78.0%). Most frequent age groups by decade were 30-39 years old (45.4%) and 40-49 years old (31.5%); sex was evenly distributed among subjects. Eight (7.1%) subjects have documented diabetes. A high prevalence of chronic blood-borne infections was observed with documented rates of HCV, HIV, and hepatitis B surface antigen (HBsAg) of 85.4% (n = 252), 27.8% (n = 82), and 7.1% (n = 21), respectively. HIV serostatus at the time of ED visit could not be determined for 10.2% (n = 30) of subjects. Among the 82 with HIV, 71 subjects had information on CD4 (median: 248, range: 2-1033) and 26 (36.6%) subjects had CD4 cell counts <200 cells per cubic millimeter [i.e., Centers for Disease Control and Prevention (CDC)-defined AIDS]. HIV RNA viral load data were available in 46 subjects, with median viral load of 60,705 copies per milliliter [range: <200 (undetectable) to >750,000; >100,000 copies/mL: 15 subjects]. Of note, 10 (4.5%) subjects were newly diagnosed with HIV infection during their visits among 223 subjects who did not have known HIV infection before the visit. Among the newly diagnosed subjects, 8 had information on CD4 cell counts, with median CD4 of 294 cells per cubic millimeter (range: 19-935; <200 cells/mm3: 25%). Four subjects had viral load information with a median viral load of 144,641 copies per milliliter (range: 8697-258,720; >100,000 copies/mL: 75%).

Of the 295 with SSTIs, specific discharge diagnoses were as follows: cellulitis only (n = 143, 48.5%), abscesses only (n = 113, 38.3%), and both (n = 39, 13.2%). None were necrotizing form of SSTIs. The leading site of involvement was right upper extremity (n = 72, 24.4%), followed by left lower extremity (n = 59, 20.0%), left upper extremity (n = 52, 17.6%), right lower extremity (n = 32, 10.9%), bilateral upper extremities (n = 15, 5.1%), and bilateral lower extremities (n = 14, 4.8%).

Back to Top | Article Outline

Bacteriology of SST and Blood Culture Results

All subjects had at least 2 blood cultures performed in the ED; 36.6% (108/295) had SST cultures performed. Rates of culture positivity were 23.1% (68/295) for blood culture and 84.3% (91/108) for SST cultures. Twenty-two (7.5%) subjects had both SST culture- and blood culture-positive results.

Microbiological findings were as follows: Among those with positive SST cultures, methicillin-susceptible Staphylococcus aureus (MSSA) (n = 36, 23.1%) was the leading pathogen in a total of 156 isolates, followed by group A streptococcus (GAS) (n = 19, 12.2%), Streptococcus anginosus (n = 18, 11.5%), MRSA (n = 15, 9.6%), and viridans group streptococci (n = 10, 6.4%). Among those with bacteremia, MSSA (n = 15, 18.8%) and GAS (n = 11, 13.8%) were the leading pathogens identified in a total of 80 isolates. Streptococcus-staphylococcus coinfections were reported in 17.6% (n = 16) of SST culture-positive subjects.

Twenty-two subjects had both positive SST and blood culture results. Nine (40.9%) had the same organism identified, with MSSA found in 7 subjects and GAS found in the other 2 subjects. Among the 13 subjects (60.1%) with discordant SST culture- and blood culture-positive results, the leading organisms identified in SST cultures were MSSA (n = 5) and GAS (n = 5), and the leading blood organisms were Propionibacterium acnes (n = 3) and Corynebacterium species (n = 3).

By specific SSTI discharge diagnosis, MSSA (n = 9, 6.3%) was the leading pathogen among 143 cellulitis only subjects, followed by GAS (n = 5, 3.5%), group C streptococcus (n = 3, 2.1%), and S. anginosus (n = 3, 2.1%). MSSA (n = 23, 20.4%) was also the leading pathogen among 113 abscesses only subjects, followed by GAS (n = 12, 10.6%), MRSA (n = 11, 9.7%), S. anginosus (n = 11, 9.7%), viridans group streptococci (n = 10, 8.8%), and group F streptococcus (n = 7, 6.2%). Finally, among 39 subjects with both cellulitis and abscesses, MSSA (n = 4, 10.3%) was the leading pathogen, followed by MRSA (n = 3, 7.7%), GAS (n = 2, 5.1%), group G streptococcus (n = 2, 5.1%), and S. anginosus (n = 2, 5.1%).

A temporal change in rates of SST MRSA was observed over the study period with only 1/26 (3.8%) S. aureus isolates MRSA positive before 2001 versus 14/25 (56%) after 2001 (P < 0.001).

Back to Top | Article Outline

Factors Associated With Increased Duration of Stay

Duration of hospital stay for participants ranged from 1 to 80 days with a median LOS of 3 days. Approximately one third of patients (34.2%, n = 101) had LOS greater than median LOS. The 7 subjects who left against medical advice were excluded from analyses of predictors for increased LOS. In bivariate analyses, being 40 years and older or having positive blood or SST culture results were associated with increased LOS, whereas having positive MRSA SST culture, type of SSTIs, HBV or HCV chronic infection, diabetes, or known HIV did not correlate with increased LOS. Having known HIV was slightly overrepresented in subjects, with duration of stay greater than median LOS. In multivariate analysis, being 40 years and older and having positive blood or SST culture were identified as predictors [hazard ratios and 95% confidence intervals as follows: bacteremia: 2.28 (1.64 to 3.17); positive SST culture: 1.55 (1.20 to 2.01); and 40 years and older: 1.31 (1.03 to 1.67)]. Further stratified analysis showed that HIV infection or SST culture positivity was associated with increased LOS only in those patients without bacteremia, but not in patients with bacteremia, with adjustment for age and type of SSTIs (Table 1).



Back to Top | Article Outline

Rehospitalization due to Febrile Illness or SSTIs Within 90 Days After Discharge

Forty-five (15.3%) IDUs with SSTIs were readmitted with febrile illness within 90 days after discharge. Among these, 44.4% (20/45) were rehospitalized with a discharge diagnosis of SSTIs. Clinical and microbiological descriptions of these 20 subjects readmitted with SSTIs are shown in Table 2. Survival analysis using Kaplan-Meier estimation demonstrated differential rates of rehospitalization by HIV infection status (Figs. 1A, B). Positivity of blood culture, the clinical duration or type of SSTIs, HBV or HCV chronic infection, and diabetes did not correlate with rehospitalization. Readmission rates for febrile illness or SSTIs in documented HIV-positive subjects were significantly higher than those subjects without known HIV. Subjects with HIV infection were approximately 3 times more likely to be readmitted due to febrile illness or SSTIs within 90 days after discharge after adjusting for age, SST culture positivity, type of SSTIs, and time to rehospitalization (Table 3). Interestingly, a higher rehospitalization rate was observed in those with newly diagnosed HIV infection than those with known infection (febrile illness: 40.0% vs 26.4%, P = 0.455; SSTIs: 20.0% vs 12.5%, P = 0.617). Of note, subjects with positive SST cultures found during initial hospitalization were 88% less likely to be readmitted due to SSTIs within 90 days (Table 3).







Back to Top | Article Outline


In this 6-year study of febrile IDUs presenting to an inner-city ED, we found SSTIs to be the leading clinical diagnosis, consistent with reports from previous studies.3,4 Among those with SSTIs, increased LOS (a proxy for increased health care resource utilization) was associated with positive culture results (either blood or SST) and being older. For those without bacteremia, positive SST culture results and being HIV positive predicted increased LOS. Furthermore, having HIV infection was associated with an increased rate of hospital readmission (another proxy for health care resource utilization).

Previous studies have reported variable findings regarding the impact of being HIV infected on clinical outcomes for those with other infectious diseases. For those studies which have stratified subjects according to CD4+ cell count, worse clinical outcomes seemed to be directly correlated with lower CD4+ counts.29-31 Other studies that have solely considered HIV infected (versus not) reported mixed results, with some reporting no differences32-39 and others showing a positive association by HIV seropositivity, with regard to LOS and mortality.40-44 The one study, which specifically evaluated the impact of HIV serostatus in drug users with SSTIs, was a relatively small retrospective chart review and reported that being HIV infected increased the likelihood of hospital admission and longer median inpatient LOS.42 The findings in our larger prospective study are consistent with and substantiated that report.

The observation that substantial numbers of newly diagnosed HIV infection are identified and that HIV seems to be a strong predictor or marker for increased health care resource utilization in febrile IDUs with SSTIs has important clinical implications for routine practice. Determining the presence of underlying HIV infection in an IDU patient whose primary presentation is an SSTI seems to be an important additional piece of clinical data that warrants consideration in determining a patient's risk for adverse outcome (ie, longer LOS or need for rehospitalization). This observation is particularly timely in light of the recently released CDC guidelines, which recommend routine HIV testing in persons aged 13-64 years in all health care settings with particular emphasis on EDs.45 The rationale for the change is earlier detection to get patient into care and for the benefits in HIV prevention. Our findings support this testing policy, and availability of the rapid bedside test and cost benefit of rapid HIV testing in acute care settings46,47 further support this practice.

Interestingly, patients were marginally significantly more likely to be readmitted with SSTIs, if they did not have positive SST culture results (ie, either negative or not done) on initial admission, even after adjusting for other covariates. A possible explanation for this finding is that those with positive SST culture results received longer duration or more appropriately tailored antibiotic therapy or more emphatic risk reduction education from providers. Although several review articles suggested an exceedingly low rate of positive blood culture in IDUs with SSTIs,15,17 at least 4 individual studies suggested much higher rates for both blood and SST culture,15,17,48,49 similar to what we report here (ie, 23% positive blood culture and 84% positive SST culture). In concordance with findings from other studies, we found S. aureus (MSSA and MRSA) and streptococci (group A, C, F, and G; viridans; and anginosus) to be the major etiologic agents responsible for SSTIs.48,50 SST culture contamination rate was very low (4 of 156, 2.6%) in our study. Four possible SST culture contaminants included Bacillus species, Corynebacterium species, Micrococcus species, and Lactobacillus species. Notably, S. aureus or streptococci were also isolated in the SST cultures of the 4 patients in whom contaminants were found. Accordingly, culture contaminants did not impact predictive effect of positive SST culture result on readmission. Our bacteriologic data do not reflect the more recently reported national trend of increased rates of MRSA infections in those with SSTIs,19,21 although MRSA accounted for more isolates after 2001. Our findings emphasize the importance of obtaining cultures both for directing therapy and tracking epidemiologic trends in drug resistance. Even though we identified a marginal reverse association between positive SST culture results at initial encounter and subsequent rehospitalization with SSTIs, further studies are needed to explore this association.

Extensive research has been conducted to identify predictors of outcome (leading to derivation and validation of decision guidelines) for patients presenting to EDs for a variety of acute infectious diseases,51-54 with pneumonia being the most well studied to date. In contrast, there have been strikingly few studies which have evaluated predictors for outcome among patients with SSTIs. Our data shed light on several key determinants which could be used to assist in derivation of hospital admission guidelines for febrile IDUs with SSTIs, the primary cause of illness and hospitalization in IDUs.

Several limitations exist. First, the study data for this investigation were drawn from a larger investigation not explicitly designed to identify determinants of increased LOS and readmission in patients with SSTIs. Accordingly, we were not able to evaluate all potential confounders (such as the size of SSTIs, detailed comorbidities, antimicrobial treatment regimens, surgery or not, frequency of IDU, and adherence with antibiotics therapy). Furthermore, significant numbers of subjects had missing information for certain variables (eg, HIV serostatus for all subjects and CD4 cell counts on all HIV-infected subjects). Despite these limitations, we were able to identify several statistically significant variables, such as HIV serostatus, which were associated with the outcome of interest; these variables warranted a more comprehensive prospective evaluation. Second, our study protocol did not allow us to determine whether patients had presented with febrile illness or SSTI-related readmissions to other institutions; therefore, our reported rates of readmission were likely underestimated. There is no a priori reason to expect, however, that rehospitalization at other institutions would be differentially distributed by patients' HIV infection status or SST culture results. Third, our study may be susceptible to enrollment selection bias. Over 25% of febrile IDUs declined to participate, and patients who received their regular care at the HIV specialty service (generally those with later stage HIV) were excluded. Although we found no significant differences in demographics between participants and nonparticipants, we were still unable to ascertain whether differences existed according to other variables (such as HIV status or rates of blood or SST culture positivity). However, even with the exclusion of those who received regular care at the HIV specialty service (who generally would be expected to have later stage HIV infection), we still found significant associations between HIV-positive serostatus and readmission, suggesting that this association likely holds true. Fourth, because one of our inclusion criteria is fever, enrollees likely had more severe SSTIs and higher rates of bacteremia than all comers with SSTIs. Thus, our findings only apply to this subgroup.

In summary, our data show that SSTIs were the primary identifiable infection among IDUs who present with febrile illnesses to EDs. The distribution of pathogens was similar to those previously reported with S. aureus and Streptococcus species predominating, but of note, after 2001, rates of MRSA rose significantly. In those with SSTIs, bacteremia was the strongest predictor for longer LOS. In those who were nonbacteremic, HIV infection was associated with increased LOS and approximately tripled the risk for rehospitalization within 90 days. This study provides pilot data for further examination of risk factors associated with adverse outcomes (ie, increased LOS and rehospitalization) for febrile IDUs with SSTIs. Future studies should be directed toward derivation and validation of ED and inpatient disposition guidelines (focusing on HIV infection and culture practice as important predictor variables), which may lead to improved outcomes for this vulnerable high-risk population.

Back to Top | Article Outline


The authors would like to thank Ambreen Khalil, Sajid Wazir, Maulik Majmudar, Marson Tenoso, and Daniel Moring-Parris for technical assistance in data collection.

Back to Top | Article Outline


1. United Nations Office on Drugs and Crime. World Drug Report 2004. Volume I. Analysis. New York, NY: United Nations Publications; 2004.
2. Kane B. Controlling disease transmission in injection drug users. Ann Intern Med. 1999;130:541-544.
3. Gordon R, Lowy F. Bacterial infections in drug users. N Engl J Med. 2005;353:1945-1954.
4. Palepu A, Tyndall M, Leon H, et al. Hospital utilization and costs in a cohort of injection drug users. CMAJ. 2001;165:415-420.
5. Binswanger I, Kral A, Bluthenthal R, et al. High prevalence of abscesses and cellulitis among community-recruited injection drug users in San Francisco. Clin Infect Dis. 2000;30:579-581.
6. Lloyd-Smith E, Kerr T, Hogg R, et al. Prevalence and correlates of abscesses among a cohort of injection drug users. Harm Reduct J. 2005;10:24.
7. Spijkerman I, van Ameijden E, Mientjes G, et al. Human immunodeficiency virus infection and other risk factors for skin abscesses and endocarditis among injection drug users. J Clin Epidemiol. 1996;49:1149-1154.
8. Kerr T, Wood E, Grafstein E, et al. High rates of primary care and emergency department use among injection drug users in Vancouver. J Public Health (Oxf). 2005;27:62-66.
9. Kimura A, Higa J, Levin R, et al. Outbreak of necrotizing fasciitis due to Clostridium sordellii among black-tar heroin users. Clin Infect Dis. 2004;38:e87-e91.
10. National Drug Intelligence Center. National Drug Threat Assessment 2003-Heroin. Available at: Accessed April 14, 2008.
11. Ciccarone D, Bourgois P. Explaining the geographical variation of HIV among injection drug users in the United States. Subst Use Misuse 2003;38:2049-2063.
12. Heinzerling K, Etzioni D, Hurley B, et al. Hospital utilization for injection drug use-related soft tissue infections in urban versus rural counties in California. J Urban Health. 2006;83:176-181.
13. Takahashi T, Baernstein A, Binswanger I, et al. Predictors of hospitalization for injection drug users seeking care for soft tissue infections. J Gen Intern Med. 2007;22:382-388.
14. Harris H, Young D. Care of injection drug users with soft tissue infections in San Francisco, California. Arch Surg. 2002;137:1217-1222.
15. Ebright J, Pieper B. Skin and soft tissue infections in injection drug users. Infect Dis Clin North Am. 2002;16:697-712.
16. Swartz M. Clinical practice. Cellulitis. N Engl J Med. 2004;350:904-912.
17. Brown P, Ebright J. Skin and soft tissue infections in injection drug users. Curr Infect Dis Rep. 2002;4:415-419.
18. Allison D, Miller T, Holtom P, et al. Microbiology of upper extremity soft tissue abscesses in injecting drug abusers. Clin Orthop Relat Res. 2007;461:9-13.
19. Frazee B, Lynn J, Charlebois E, et al. High prevalence of methicillin-resistant Staphylococcus aureus in emergency department skin and soft tissue infections. Ann Emerg Med. 2005;45:311-320.
20. Fridkin S, Hageman J, Morrison M, et al. Methicillin-resistant Staphylococcus aureus disease in three communities. N Engl J Med. 2005;352:1436-1444.
21. Moran G, Krishnadasan A, Gorwitz R, et al. Methicillin-resistant S. aureus infections among patients in the emergency department. N Engl J Med. 2006;355:666-674.
22. Centers for Disease Control and Prevention (CDC). Community-acquired methicillin-resistant Staphylococcus aureus infections: Michigan. MMWR Morb Mortal Wkly Rep. 1981;30:185-187.
23. Saravolatz L, Markowitz N, Arking L, et al. Methicillin-resistant Staphylococcus aureus. Epidemiologic observations during a community-acquired outbreak. Ann Intern Med. 1982;96:11-16.
24. Craven D, Rixinger A, Goularte T, et al. Methicillin-resistant Staphylococcus aureus bacteremia linked to intravenous drug abusers using a “shooting gallery”. Am J Med. 1986;80:770-776.
25. Fleisch F, Zbinden R, Vanoli C, et al. Epidemic spread of a single clone of methicillin-resistant Staphylococcus aureus among injection drug users in Zurich, Switzerland. Clin Infect Dis. 2001;32:581-586.
26. Hsieh Y, Majmudar M, Khalil A, et al. Epidemiologic, clinical features, and 90 day outcomes of infective endocarditis in febrile injection drug users. Clin Res Cardiol. 2007;96:428.
27. Kirchhoff L, Sheagren J. Epidemiology and clinical significance of blood cultures positive for coagulase-negative staphylococcus. Infect Control. 1985;6:479-486.
28. Cox D. Regression models and life-tables. J R Stat Soc. 1972;34:187-220.
29. Holmes C, Losina E, Walensky R, et al. Review of human immunodeficiency virus type 1-related opportunistic infections in sub-Saharan Africa. Clin Infect Dis. 2003;36:652-662.
30. Harms G, Feldmeier H. The impact of HIV infection on tropical diseases. Infect Dis Clin North Am. 2005;19:121-135.
31. Feldman C. Pneumonia associated with HIV infection. Curr Opin Infect Dis. 2005;18:165-170.
32. Jones N, Huebner R, Khoosal M, et al. The impact of HIV on Streptococcus pneumoniae bacteraemia in a South African population. AIDS. 1998;12:2177-2184.
33. Ribera E, Miro J, Cortes E, et al. Influence of human immunodeficiency virus 1 infection and degree of immunosuppression in the clinical characteristics and outcome of infective endocarditis in intravenous drug users. Arch Intern Med. 1998;158:2043-2050.
34. Frankel R, Virata M, Hardalo C, et al. Invasive pneumococcal disease: clinical features, serotypes, and antimicrobial resistance patterns in cases involving patients with and without human immunodeficiency virus infection. Clin Infect Dis. 1996;23:577-584.
35. Pulvirenti J, Kerns E, Benson C, et al. Infective endocarditis in injection drug users: importance of human immunodeficiency virus serostatus and degree of immunosuppression. Clin Infect Dis. 1996;22:40-45.
36. Christensen D, Feldman C, Rossi P, et al. HIV infection does not influence clinical outcomes in hospitalized patients with bacterial community-acquired pneumonia: results from the CAPO international cohort study. Clin Infect Dis. 2005;41:554-556.
37. Bhagwanjee S, Muckart D, Jeena P, et al. Does HIV status influence the outcome of patients admitted to a surgical intensive care unit? A prospective double blind study. BMJ. 1997;314:1077-1081.
38. Feldman C, Glatthaar M, Morar R, et al. Bacteremic pneumococcal pneumonia in HIV-seropositive and HIV-seronegative adults. Chest. 1999;116:107-114.
39. Cacala S, Mafana E, Thomson S, et al. Prevalence of HIV status and CD4 counts in a surgical cohort: their relationship to clinical outcome. Ann R Coll Surg Engl. 2006;88:46-51.
40. Mugo N, Kiehlbauch J, Nguti R, et al. Effect of human immunodeficiency virus-1 infection on treatment outcome of acute salpingitis. Obstet Gynecol. 2006;107:807-812.
41. Kwara A, Roahen-Harrison S, Prystowsky E, et al. Manifestations and outcome of extra-pulmonary tuberculosis: impact of human immunodeficiency virus co-infection. Int J Tuberc Lung Dis. 2005;9:485-493.
42. Scheidegger C, Zimmerli W. Incidence and spectrum of severe medical complications among hospitalized HIV-seronegative and HIV-seropositive narcotic drug users. AIDS. 1996;10:1407-1414.
43. Johnson D, Carriere K, Houston S, et al. Hospitalization for community-acquired pneumonia in Alberta patients with human immunodeficiency virus infections: a case control study. Can Respir J. 2003;10:265-270.
44. Mrus J, Braun L, Yi M, et al. Impact of HIV/AIDS on care and outcomes of severe sepsis. Crit Care. 2005;9:R623-R630.
45. Branson B, Handsfield H, Lampe M, et al. Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings. MMWR Recomm Rep. 2006;55:1-17.
46. Lubelchek R, Kroc K, Hota B, et al. The role of rapid vs conventional human immunodeficiency virus testing for inpatients: effects on quality of care. Arch Intern Med. 2005;165:1956-1960.
47. Lyss S, Branson B, Kroc K, et al. Detecting unsuspected HIV infection with a rapid whole-blood HIV test in an urban emergency department. J Acquir Immune Defic Syndr. 2007;44:435-442.
48. Chen J, Fullerton K, Flynn N. Necrotizing fasciitis associated with injection drug use. Clin Infect Dis. 2001;33:6-15.
49. Takahashi T, Merrill J, Boyko E, et al. Type and location of injection drug use-related soft tissue infections predict hospitalization. J Urban Health. 2003;80:127-136.
50. Summanen P, Talan D, Strong C, et al. Bacteriology of skin and soft-tissue infections: comparison of infections in intravenous drug users and individuals with no history of intravenous drug use. Clin Infect Dis. 1995;20:S279-S282.
51. Gaeta T, Webheh W, Yazji M, et al. Respiratory isolation of patients with suspected pulmonary tuberculosis in an inner-city hospital. Acad Emerg Med. 1997;4:138-141.
52. Corwin P, Toop L, McGeoch G, et al. Randomised controlled trial of intravenous antibiotic treatment for cellulitis at home compared with hospital. BMJ. 2005;330:129.
53. Nguyen H, Corbett S, Menes K, et al. Early goal-directed therapy, corticosteroid, and recombinant human activated protein C for the treatment of severe sepsis and septic shock in the emergency department. Acad Emerg Med. 2006;13:109-113.
54. Riley P, Aronsky D, Dean N. Validation of the 2001 American Thoracic Society criteria for severe community-acquired pneumonia. Crit Care Med. 2004;32:2398-2402.

bacteriology; HIV; injection drug users (IDUs); length of stay (LOS); rehospitalization; skin/soft tissue infections (SSTIs)

© 2008 Lippincott Williams & Wilkins, Inc.