Thomas Lodise, PharmD, PhD

infection (CDI) is a major and urgent threat that remains one of the most common health care-associated infections in US hospitals, affecting approximately half a million people annually.

 Caused by a contagious bacterium of the same name, CDI can become a serious health threat to not only patients with CDI, but also others in the home and the community.

During the COVID-19 pandemic, there has been a call for renewed attention on CDI, particularly occurrences of CDI with recurrent infection. This increase in attention may be a result of the increase in antibiotics being prescribed to patients with COVID-19 to treat or prevent bacterial infections such as pneumonia—antibiotic use can be a risk factor for CDI.

Since the pandemic began, an estimated 72% of patients with COVID-19 were treated with broad-spectrum antibiotics.

 With this increase antibiotic treatment, pharmacists can play an important role in educating patients and other health care practitioners on the signs and symptoms of CDI, as well as the risk factors associated with antibiotic use.

1. Although CDI can affect anyone, the risk is higher for certain populations.

Those at higher risk are people who are taking, or have recently taken, antibiotics, who have spent some time in a hospital or in a long-term care facility, who have a weakened immune system, or who are aged 65 years or older.

The most common symptoms for CDI include diarrhea, nausea, stomach pain or cramps.

 If a patient complains about these symptoms over a prolonged period of time, especially after antibiotic use, a pharmacist should encourage the patient to alert their physician as soon as possible.

If not treated right away, CDI can lead to a very serious acute medical condition and could even be fatal, especially in highly vulnerable patient populations (ie, patients ≥ 65 years of age, patients who reside in long-term care facilities, patients with weakened immune systems, patients with multiple comorbidities, etc).

2. CDI is contagious and can be spread to others who may develop a serious infection.

CDI can be spread by direct contact, indirect contact, or by contact with contaminated surfaces.

 However, CDI is typically transmitted from person to person by the fecal-oral route.

Pharmacists can educate patients who have been diagnosed with CDI and those experiencing symptoms on the importance of good hygiene. Regular handwashing with soap and ensuring surfaces are washed down with a strong cleaner are keys to the prevention and spread of 

3. Antibiotics are a predominant risk factor for CDI and its recurrence

Pharmacists can help educate other health care practitioners that use of antibiotics for extended periods of time, or the use of more than 1 antibiotic for treating an illness, can raise the risk of CDI

The disruption of the intestinal microbiota by antibiotics is long-lasting, and risk of CDI increases both during therapy and in the 3-month period following cessation of therapy.

While the risk of CDI increases as a function of antibiotic duration and number of antibiotics received, even 1 to 2 days of antibiotics can increase a patient’s risk of 

Treatment of CDI is extremely challenging and patients with CDI often require multiple CDI treatment courses as recurrences occur in up to 35% of people who get CDI after the initial diagnosis.

 After the first recurrence, it has been estimated that up to 60% of patients are likely to get a subsequent recurrence.

restoring a healthy gut microbiome is increasingly accepted as a promising treatment option.

In recent years, scientists have learned a great deal about the role of the gut microbiome, which is the microbial community in the intestinal tract which influences metabolism and immunity and acts as a mediator of resistance to some pathogenic infections. Composed of more than 100 trillion bacteria, the gut microbiome has the capacity to promote overall health.

Disruption of the gut microbiome (also known as dysbiosis) leads to an intestinal microenvironment susceptible to pathogenic insult from opportunistic bacteria such as 

. While antibiotics are the standard of care for CDI, they do not address the underlying dysbiosis and play a large role in the vicious cycle of recurrence.

 The aim of microbiome restoration is to repopulate the diverse gut microbiota to treat the disease.

5. Therapeutic options for gut microbiome restoration are limited and in need of an update.

For recurrent CDI, one historic approach to microbiome restoration has been fecal microbiota transplant (FMT). Data suggest FMT is efficacious for the treatment of recurrent CDIand reestablishes biodiversity in the gut. However, variability exists across clinical studies—cure rates were lower in randomized controlled trials than in open-label studies (67.7% vs 82.7%, respectively; P<.001).

Likewise, the lack of product standardization and administration methods has created a situation where a regulated, safe, and effective product is critically needed.

 Researchers are exploring promising microbiome-based therapeutics that work to restore a healthy, diverse mix of gut bacteria to evolve standards of care for recurrent CDI.

 is a professor at Albany College of Pharmacy and Health Sciences.

Centers for Disease Control and Prevention website. 2021 Antibiotic Resistance Threats Report: 

. https://www.cdc.gov/drugresistance/pdf/threats-report/clostridioides-difficile-508.pdf. Accessed December 6, 2021.

Guh AY, Mu Y, Winston LG, et al. Trends in U.S. burden of 

infection and outcomes. N Engl J Med. 2021;382(14):1320-1330.

 infection. N Engl J Med. 2015;372(16):1539-1548.

Fernandez-Garcia L, Blasco L, Lopez M, Tomas M. Clostridium difficle infection: pathogenesis, diagnosis and treatment, Clostridium 

– A Comprehensive Overview, Shymaa Enany, Intech Open, DOI: 10.5772/67754. Available from: https://www.intechopen.com/chapters/54496

Spigaglia P. (2020). COVID-19 and Clostridioides

 infection (CDI): Possible implications for elderly patients. 

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7315154/

in COVID-19 patients. Emerg Infect Dis. 26(9):2020.

Bien J, Palagani V, Bozko P. The intestinal microbiota dysbiosis and Clostridium 

 infection: is there a relationship with inflammatory bowel disease? Therap Adv Gastroenterol. 2013;6(1):53-68.

Centers for Disease Control and Prevention. What Is C. Diff? 17 Dec. 2018. Available at: 

Surawicz CM, Brandt LJ, Binion DG, et al. Guidelines for diagnosis, treatment , and prevention of Clostridium 

 infection. Am J Gastroenterol. 2013;108(4):478-498.

McDonald LC, Gerding DN, Johnson S, et al. Clinical Practice Guidelines for Clostridium 

 Infection in Adults and Children: 2017 Update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA). Clin Infect Dis. 2018;66(7):e1-e48.

Johnson S, Lavergne V, Skinner AM, et al. Clinical Practice Guidelines by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA): 2021 Focused Update Guidelines on Management of Clostridioides

Infection in Adults. Clin Infect Dis. 2021;73(5):e1029-e1044.

Langdon A, Crook N, Dantas G. The effects of antibiotics on the microbiome throughout development and alternative approaches for therapeutic modulation. Genome Med. 2016;8(1):39.

Kao D, Roach B, Silva M, et al. Effect of oral capsule- vs colonoscopy-delivered fecal microbiota transplantation on recurrent Clostridium 

 infection: a randomized clinical trial. JAMA. 2017;318(20):1985-1993

Rodrigues R, Barber GE, Ananthakrishnan AN. A comprehensive study of costs associated with recurrent Clostridioides 

 infection. Infect Control Hosp Epidemiol. 2017;38(2):196-202

Unni S, Scott T, Boules M, Teigland, C, Parente A, Nelson W. Healthcare burden and costs of recurrent Clostridioides 

infection in the Medicare population. Presented at: AMCP 2020; April 21-24, 2020; Houston, Tx.

Lurienne L, Bandinelli P, Galvain T, Coursel CA, Oneto C, Feuerstadt P. Perception of quality of life in people experiencing or having experienced a Clostridioides 

 infection: a US population survey. J Patient Rep Outcomes. 2020;4(1):14.

van Nood E, Vrieze A, Nieuwdorp M, et al. Duodenal infusion of donor feces for recurrent Clostridium 

Joseph J, Saha S, Greenberg-Worisek AJ. Fecal microbiota transplantation: an ambiguous translational pathway for a promising treatment. Clin Transl Sci. 2019;12(3):206-208.

Hensgens MP, Goorhuis A, Dekkers OM, Kuijper EJ. Time interval of increased risk for Clostridium

 infection after exposure to antibiotics. J Antimicrob Chemother 2012(67):742-8.

Pepin J, Saheb N, Coulombe MA, et al. Emergence of fluoroquinolones as the predominant risk factor for Clostridium 

-associated diarrhea: a cohort study during an epidemic in Quebec. Clin Infect Dis 2005(41):1254-1260.

Articles

Pharmacists play an important role in educating patients and other health care practitioners on the signs and symptoms of C. difficile infection.

Five Things Pharmacists Should Know About C. difficile Infection

Complicated skin and skin structure infections (cSSSIs are infections that involve deeper layers of the skin, including the muscle and fascia. In the United States, there are an estimated 14 million outpatient (ie, physician offices, emergency and outpatient departments) healthcare visits for suspected

 Treatment of cSSSIs is complicated and often involves both surgery and intravenous antibiotics.

Once considered a disease predominated by 

, it is now estimated that more than half of all cSSSIs are due to Staphylococcus aureus, many of which are methicillin  resistant.

 In a study of 11 emergency departments in the United States, approximately 76% of purulent (ie, containing pus) skin and soft tissue infections in adults were caused by 

.10 Of these infections, 78% were caused by methicillin-resistant 

 (MRSA); overall, MRSA caused 59% of skin and soft tissue infections. Given the increasing ubiquity of cSSSIs due to MRSA, 2 commonly prescribed empiric antibiotics are vancomycin and linezolid.

Both linezolid and vancomycin are approved by the US Food and Drug Administration (FDA) for the treatment of cSSSIs caused by Gram-positive bacteria. These antibiotics are approved for patients with similar clinical characteristics, including patients with deep soft tissue infections, surgical/traumatic wound infections, major abscesses, cellulitis, and infected ulcers and burns. The Infectious Diseases Society of America practice guidelines recommend both intravenous vancomycin and oral or intravenous linezolid 600 mg twice daily as first-line therapy for cSSSIs.

 To date, vancomycin and linezolid have been evaluated in a number of comparator-controlled cSSSI trials. Overall, the efficacy and safety of vancomycin versus linezolid for cSSSI due to MRSA have been comparable.

 Most recently, similar clinical and microbiologic success rates were found in a comparative clinical trial that assessed linezolid versus vancomycin for the management of noncellulitis cSSSI due to MRSA.

While clinical trials have established the safety and efficacy of linezolid relative to vancomycin for cSSSI, many important questions remain. Patients enrolled in clinical trials may not be fully reflective of the diverse patient populations that use the drugs in clinical practice. Therefore, practice-level analyses are needed to ascertain the real-world effectiveness of linezolid relative to vancomycin in patients with cSSSIs. In addition, the primary outcomes assessed in cSSSI clinical trials center on clinical and microbiologic  success rates at predefined times such as 1 to 2 weeks after the completion of therapy (ie, test-of-cure visit). To fully understand the clinical and economic implications of one therapy relative to another for cSSSIs, it is necessary to make a comprehensive  assessment of real-world outcomes including healthcare utilization and rehospitalization and costs, in time frames beyond those typically seen in clinical trials. Such a comprehensive assessment, which examines healthcare utilization and costs across  treatment groups, informs the decisionmaking processes of public and private insurers/payers and ultimately leads to increased patient health/utility and lower overall medical costs for the patient or society.

Our previous study reported that patients who were treated with linezolid versus vancomycin after hospital discharge had lower rates of repeat hospitalization across an array of infection types; however, the prior study did not provide cost savings estimates.

The current study aims to build upon our prior work using updated data from 2 administrative claims data sets (HealthCore Integrated Research Database and IMS PharMetrics database). This time, we examined both rehospitalization rates and total direct medical costs for patients treated with linezolid and patients treated with vancomycin following hospital discharge for a cSSSI from the perspective of a commercial insurer in the United States.

This was a retrospective observational, comparative effectiveness drug study (

). The study design and analysis reflected the perspective of private insurers. The cSSSI cases were identified based on eligible hospitalizations occurring between January 1, 2007, and September  30, 2009. The index date was defined as the discharge date of the qualifying hospitalization. We examined

baseline comorbidities, prior hospitalizations, and prior infections in the preindex period, which was defined as 180 days prior to the index hospitalization. A primary postindex period of 42 days (sensitivity analyses at 30 and 180 days) from the discharge date of the qualifying hospitalization was used for comparing rehospitalizations and total costs, in accordance with our prespecified analysis plan. During the stage of developing our analysis plan, 42 days was specifi ed as the primary end point for the analysis because based on the literature and clinical insight, most infections should resolve within a maximum of 6 weeks (ie, 42 days).

The study population consisted of adult patients (aged 18-64 years) with a hospitalization related to cSSSI within either the HealthCore Integrated Research Database or the IMS PharMetrics database based on 

International Classification of Diseases, Ninth Revision, Clinical Modification

) codes used to defi ne cSSSI. The codes used for the analyses are available in the  

). Inclusion criteria required that patients be treated with either linezolid or vancomycin within 7 days of their hospital discharge. The following patients were excluded: patients who were enrolled in Medicare (because of unobserved Medicare claims data); patients who received oral vancomycin (because there is no systemic absorption of oral vancomycin, it is not used for treating cSSSIs); patients who had fewer than 3 days of linezolid oral therapy, in order to focus on those patients who experienced a therapeutic effect (there was no minimum day requirement for intravenous therapy because therapeutic drug levels are achieved after 1 dose); and patients with index hospitalizations longer than 30 days. In addition, patients with infections of the musculoskeletal system (codes 015.xx, 324.1, 376.03, 711.xx, 715.90, 715.98, 719.06, 722.9, 722.93, 726.65,  727.05, 727.09, 730. xx), endocarditis (code 424.9x), meningitis (codes 320.2 320.7, 320.8x), necrotizing fasciitis (codes 728.86, 729.4), gangrene (code 785.4), or prosthetic implant/device infections (codes 919.7, 996.6x, 999.3x) during the index  hospitalization were excluded. These exclusion criteria were developed to facilitate a comparative study of linezolid versus vancomycin based on the FDA-approved linezolid indication for cSSSI.

The following outcomes were captured after discharge: all-cause rehospitalization and rehospitalization related to a cSSSI infection  and total medical costs, which included index drug (ie, linezolid or vancomycin) costs; other drug costs; and inpatient, outpatient, and other costs such as laboratory and diagnostic costs. In accordance with our prespecifi ed analysis plan, main analyses examined the outcomes at 42 days, while sensitivity analyses focused on the outcomes at 30 and 180 days following discharge. The covariate of interest was receipt of linezolid (vsvancomycin) as the index drug. The following potential confounders were  examined and controlled for in multivariable regression models: sex, age at index discharge, Charlson Comorbidity Index (CCI)  score, length of stay of the index hospitalization, presence of bacteremia during the index hospitalization, presence of sepsis during  the index hospitalization, use of other anti-MRSA antibiotics in the preindex period, presence of any infection-related hospitalization or infection-related outpatient physician visit in the preindex period, and presence of any all-cause hospitalization in the preindex period. The effect of CCI score was modeled using an ordered categorical measure (no comorbidity, 1 comorbidity, 2  comorbidities, 3 or more comorbidities) in both the rehospitalization and cost regression models. The rehospitalization and cost  regression models included a continuous (linear) measure of age. The cost regression model also included a quadratic term to  capture the nonlinear effect of age. Length of stay was dichotomized at the median value (5 days) in the rehospitalization regression model and was log-transformed in the cost regression model. All other variables entered as binary indicators of the factor described.

 value of .05 was used to determine statistical significance in hypothesis testing, and a 

 value of .10 was used in covariate selection. Descriptive analyses examined various characteristics of the study cohort. The χ

 statistics were used to test for significant differences in categorical and continuous variables, respectively, between linezolid and vancomycin users. A correlatio matrix including all study variables was used to identify potential multicollinearity problems. Multivariable logisti regression analyses were used to estimate the adjusted odds of rehospitalization for linezolid versus vancomycin users. Prior to model specification, normality tests combined with quantile-quantile plots confi rmed the suitability of the gamma distribution for the cost data. Following estimation of model parameters using a generalized linear model with a gamma distribution, the modified Park test

 confirmed the appropriateness of the gamma family distribution for this analysis. A generalized linear model with log link was used to examine differences in total direct medical costs between linezolid versus vancomycin, after controlling for covariates. The covariates from the generalized linear model with gamma distribution provided input data for calculating the marginal effect (ie, the covariate-adjusted incremental cost for linezolid users compared with vancomycin users). The 95% confidence interval (CI) on the adjusted incremental cost was defi ned using the 5th and 95th percentiles of bootstrapped replicates (1000 runs).

To address the potential for selection bias, we ran the aforementioned multivariable analyses in a propensity score—matched sample as well as in the full sample. The propensity score–matched sample was created by estimating the propensity of receiving linezolid as the index drug and then using a greedy algorithm macro written by Jon Kosanke and Erik Bergstralh of the Mayo Clinic College of Medicine to match cases with controls.

 The algorithm used a caliper width of 0.2 of the standard deviation (SD) of the logit of the propensity score, which is a function of the predicted propensity score, to conduct a 1-to-1 match. The following baseline characteristics  were candidate variables for the regression model that estimated the propensity score: sex, age, length of stay of  the index hospitalization, bacteremia during the index hospitalization, sepsis during the index hospitalization, CCI score, cancer at baseline, transplant, primary immunodefi ciency disorders, human immunodefi ciency virus, renal insufficiency, diabetes,  hospitalization in the preindex period, and anti-MRSA antibiotics use in the preindex period; a forward selection approach was used to obtain the final set of variables for the regression equation. After the propensity score—matched sample was created, we compared the distribution of the propensity score between matched linezolid and vancomycin patients, examined standardized differences in observed baseline  patient characteristics, and conducted individual variable and joint distribution significance testing.

The demographic and clinical characteristics of patients treated with linezolid versus vancomycin are shown in 

. Of the 7260 patients with a cSSSI-related hospitalization in the full sample, 43% (n = 3093) initially received linezolid and 57% (n = 4167) initially received vancomycin after discharge. While patient characteristics were similar, those treated with vancomycin after discharge had more baseline comorbidities, longer hospital stays, and higher rates of bacteremia during the index hospitalization relative to the linezolid users. The propensity score—matched sample consisted of 5920 patients with equal numbers in each treatment group and with no significant differences in examined baseline characteristics (

Results of bivariate rehospitalization comparisons are displayed in 

. Unadjusted rehospitalization rates were lower for linezolid users compared with vancomycin users at each postdischarge time point assessed. At 42 days after discharge, treatment with linezolid versus vancomycin was correlated with signifi cantly lower rates of all-cause rehospitalizations (13% vs 22%; 

 <.01) and cSSSI-related rehospitalizations (8% vs 15%; 

 <.01). Similarly, compared with vancomycin, linezolid use was associated with lower rates of all-cause and cSSSI-related rehospitalizations at day 30 (11% vs 19% and 7% vs 13%) and day 180 (24% vs 36% and 12% vs 21%), respectively. The differences were statistically signifi cant in all cases (

Results of the unadjusted mean costs by cost category for patients treated with linezolid versus vancomycin are displayed in 

. Index drug costs at 42 days after discharge were signifi cantly higher for those treated with linezolid versus vancomycin ($1634 vs $375;

 <.01); however, unadjusted total direct medical costs, inclusive of index drug costs, were significantly lower among patients treated with linezolid versus vancomycin ($6258 vs $10,066; 

 <.01). Similar results were found at 30 and 180 days after discharge (Figure 3).

After controlling for covariates (age, sex, CCI score, bacteremia, sepsis, length of stay during the index hospitalization, hospitalization or use of other anti-MRSA antibiotics in the preindex period), the use of linezolid versus vancomycin was associated with a lower odds of rehospitalization; linezolid use was associated with 40% lower odds of rehospitalization for any reason within 42 days of initial discharge (odds ratio [OR] = 0.60, 95% CI, 0.53-0.69; 

, full sample).  Additional factors correlated with the odds of all-cause rehospitalization within 42 days included more comorbidities at baseline (OR = 1.35, 95% CI, 1.14-1.60 for CCI score of 1; OR = 1.95, 95% CI, 1.59-2.38 for CCI score of 2; OR = 2.55, 95% CI, 2.16-3.01 for CCI score of >3); having an index hospitalization longer than the median of 5 days (OR = 1.30, 95% CI, 1.13-1.48); and having an  all-cause hospitalization in the preindex period (OR = 1.68, 95% CI, 1.47-1.90). The results were similar when examining all-cause rehospitalizations for linezolid versus vancomycin within 30 days (OR = 0.58, 95% CI, 0.50-0.67) or 180 days after discharge (OR = 0.65, 95% CI, 0.58-0.73).

Table 2 (matched sample) further documents that linezolid was associated with reduced odds of repeat hospitalization in the subset of propensity score—matched patients. After controlling for covariates, using linezolid as initial treatment was associated with a 41% lower likelihood of rehospitalization for any reason within 42 days of initial discharge (OR = 0.59, 95% CI, 0.51-0.68). The results were similar when examining all-cause rehospitalizations within 30 days (OR = 0.56, 95% CI, 0.48-0.65) or 180 days after discharge (OR = 0.66, 95% CI, 0.59-0.74).

 After controlling for covariates, the use of linezolid compared with vancomycin was associated with a 45% lower likelihood of a cSSSI-related rehospitalization within 42 days of initial discharge (OR = 0.55, 95% CI, 0.47-0.64; Table 2, full sample). Individuals with 2 or more index comorbidities (OR = 1.57, 95% CI, 1.24-2.00 for CCI score of 2; OR = 1.92, 95% CI, 1.58-2.34 for CCI score of 

3), an index hospitalization longer than the median of 5 days (OR = 1.21, 95% CI, 1.03-1.41), and any all-cause hospitalization in the preindex period (OR = 1.34, 95% CI, 1.15-1.56) had greater odds of rehospitalization with an 

code for cSSSI infection within 42 days of discharge. The likelihood of having a cSSSI-related rehospitalization

within 30 and 180 days was similar to that at 42 days after discharge. Linezolid use was associated with a 49% reduction in the odds of rehospitalization at 30 days (OR = 0.51, 95% CI, 0.43-0.61) and a 42% reduction at 180 days (OR = 0.58, 95% CI, 0.51-0.67) compared with vancomycin use following discharge after index hospitalization.

Table 2 (matched sample) shows that linezolid was associated with reduced odds of a cSSSI-related rehospitalization in the subset of propensity score—matched patients. After controlling for covariates, using linezolid as initial treatment was associated with a 45% lower likelihood of cSSSI-related rehospitalization within 42 days of initial discharge (OR = 0.55, 95% CI, 0.46-0.65). The results were similar when examining cSSSI-related rehospitalization within 30 days (OR = 0.51, 95% CI, 0.42-0.61) or 180 days after discharge (OR = 0.58, 95% CI, 0.50-0.67).

After controlling for covariates, linezolid versus vancomycin use was associated with a mean cost savings of $1420 (95% CI, $1368-$1472) per patient within 42 days of discharge in the full sample and with a mean cost savings of $1708 (95% CI, $1647-$1769) per patient in the propensity score—matched sample (Table 2). The estimated cost savings over a 30-day postindex period were $1003 (95% CI, $970-$1036) in the full sample and $1254 (95% CI, $1211-$1297) in the propensity score–matched sample. The estimated cost savings over a 180- day postindex period were $2539 (95% CI, $2436-$2641) in the full sample and $3157 (95% CI, $3031-$3284) in the propensity score–matched sample.

Several studies have evaluated the economic effect of linezolid compared with vancomycin in the treatment of skin and skin structure infections, including cSSSIs, and these are described in the medical literature.

 In these studies, linezolid  compared with vancomycin therapy was generally associated with a decrease in intravenous therapy days

 Linezolid therapy appeared to either decrease costs

 compared with vancomycinfor the treatment of cSSSIs, including those caused by MRSA.

Although clinical trials have established the efficacy of linezolid relative to vancomycin for cSSSI, there is still a major need for clinical and economic comparative data in the diverse patient populations that use these drugs in clinical practice. For cSSSIs that require hospitalization, rehospitalization rates and post-discharge costs are significant concerns from both a public health and a payer perspective. Unfortunately, these data are not well captured from cSSSI registration trials. According to the CMS, 20% of all readmissions occur within 30 days of discharge and 76% of these readmissions may be preventable.

 Beyond the clear public health implications, these potentially preventable high readmission rates are particularly concerning given the recent nosocomial infection policy initiative within the Centers for Medicare and Medicaid Services; it will refuse to pay for additional costs associated with selected nosocomial infections that are “reasonably preventable.”

In light of the need for real-world comparative effectiveness data, this study compared rehospitalization rates and total direct medical costs among commercially insured adult patients with cSSSIs treated with either linezolid or vancomycin after hospital discharge. Data were compared across 2 large national administrative claims databases. The results of the rehospitalization analyses indicated that receipt of linezolid was associated with an approximate 40% lower odds of a rehospitalization; the exact estimate of lower odds varied from 35% to 49% according to length of follow-up (30, 42, and 180 days postdischarge) and whether all-cause or cSSSI-related rehospitalizations were examined. These results are similar to our previous administrative claims analysis that pooled 21  unique infection types.

 In that study of 11,018 hospitalized patients, linezolid was associated with a lower OR of repeat infection- related hospitalization in patients with pneumonia, S aureus, and skin infections. For patients with skin infections, linezolid was associated with a reduction of 27% in the odds of infection-related hospitalization relative to vancomycin. When we examined the full sample of 21 infection types, the number needed to treat with linezolid instead of vancomycin to avoid a rehospitalization was 14. For the current study, the number needed to treat with linezolid versus vancomycin was 11 or 15 to avoid a repeat hospitalization, depending on whether the analysis considers all-cause or cSSSI-related hospitalizations.

Although rehospitalization costs were not calculated, we did examine the rates of rehospitalization and saw that linezolid use was associated with signifi cantly lower rates. The current study also showed that use of linezolid versus vancomycin as the index drug for patients hospitalized with cSSSI may be associated with significantly lower total costs of care (including rehospitalization costs) compared with postdischarge vancomycin. The current study also showed an adjusted lower total direct cost savings per patient between $1003 and $2539 for patients treated with linezolid versus patients treated with vancomycin. In the multivariable regression analysis that controlled for potential confounders, these lower total direct medical costs were consistent at 30, 42, and 180 days. These results are similar to those published by McKinnon and colleagues.

 Using data from a clinical trial comparing linezolid with vancomycin for the treatment of cSSSIs due to suspected or proven MRSA, the mean ± SD cost for intent-to-treat population patients (n = 717) treated with linezolid versus vancomycin was $4865 ± $4367 versus $5738 ± $5190, respectively (difference, $873; 

 = .02); in the MRSA population it was $4881 ± $3987 versus $6006 ± $5039, respectively (difference, $1125; 

 = .04). A separate administrative claims database study by McKinnon and colleagues showed similar findings.22 In a cSSSI subgroup, mean ± SD healthcare costs were greater for patients receiving vancomycin compared with those receiving linezolid (vancomycin, $9244 ± $14,244; linezolid, $7794 ± $16,910; 

 = .12), and the rate of relapse or rehospitalization was lower for patients receiving linezolid compared with those receiving vancomycin (OR = 0.70, 95% CI, 0.35-1.40).

The findings in this study should be interpreted in the context of a number of important limitations of observational studies. First, this study reflects nonrandom assignment based upon real-world clinical use of linezolid versus vancomycin. Thus, selection bias may impact the reported findings. Furthermore, we utilized administrative claims data, which are collected for payment purposes rather than research. Therefore, the presence of a claim for a drug does not guarantee the actual administration of the drug; the data do not capture the dispensing of a drug sample; and data may be missing if all the claims are not submitted by providers. Second, this  administrative claims database study lacked patient clinical information (eg, incision and drainage procedures), including mortality.  Instead of culture results, 

 codes were used to determine the presence of infection. Furthermore, administrative claims  data do not allow us to look at the specifi c type of infection, bacterial etiology (eg, methicillin-sensitive 

ersus MRSA), and antimicrobial susceptibility of the cultured bacteria. Third, we were not able to control for unobservable patient,  rovider, and insurance benefit design characteristics that would further impact selection bias regarding prescription of linezolid or  vancomycin, or for confounding by indication. Selection bias may be a particular issue if linezolid is prescribed to patients with  certain demographic  characteristics or higher socioeconomic status because of the greater cost and patient cost sharing for  linezolid. However, to limit the impact of selection bias, we controlled for certain observable characteristics including age, sex, CCI score, bacteremia during the index hospitalization, length of stay of the index hospitalization, and any use of other anti-MRSA   antibiotics in the preindex period. We also replicated our analyses in a subset of patients who were matched using a propensity score—matching algorithm, and the results were similar for those analyses. Finally, information regarding the antibiotics received during the index hospitalization was not available. In addition to the described limitations, it is important not to extend the findings beyond our study population of nonelderly patients with private insurance, because they are not necessarily representative of all nonelderly, privately insured patients in the United States.

The strength of this study is the large number of patients included in this pooled database (n = 7260). The data were consistent across different defi nitions of rehospitalization and total direct medical costs when evaluated at 30, 42, and 180 days following discharge. Patients prescribed linezolid after hospital discharge had signifi cantly lower rates of all-cause and cSSSI-related rehospitalization, and lower total direct costs compared with patients receiving vancomycin for cSSSI. In an era when value-based medical purchasing and reduction in wasteful spending are major considerations in health policy, such data provide valuable evidence for hospital administrators and payers. This is particularly true in infectious diseases, where avoidable readmissions may not be reimbursed.

Among commercially insured patients with a cSSSI-related  hospitalization, the use of linezolid versus vancomycin as the index drug following hospital discharge was associated with signifi cantly higher antibiotic drug costs. However, linezolid use was associated with statistically signifi cant lower total direct medical costs, including the index drug costs, compared with vancomycin use. Furthermore, linezolid was associated with significantly lower rehospitalization rates than vancomycin in both unadjusted comparisons and multivariable analyses that controlled for age, sex, CCI score, bacteremia, sepsis, length of stay during the index  hospitalization, prior hospitalization, or use of other anti-MRSA antibiotics in the preindex period. Results were consistent across the various postdischarge end points examined and across study samples (ie, full unmatched and propensity score—matched samples).

Among commercially insured patients treated with linezolid or vancomycin following cSSSI-related hospitalization, linezolid was associated with significantly lower rehospitalization rates and total direct medical costs.