Background Few data exist on the efficacy of the long-acting tetracyclines doxycycline and minocycline against methicillin-resistant Staphylococcus aureus (MRSA) infection.
Methods The medical records of 24 patients with serious tetracycline-susceptible MRSA infections who were treated with doxycycline or minocycline were reviewed. A review of the literature on the use of these antibiotics for treatment of both methicillin-susceptible and methicillin-resistant S. aureus infection was also performed.
Results Complicated skin and skin-structure infections were most common (67%). Clinical cure was achieved in 20 (83%) of 24 patients in our case series. Both drugs were well-tolerated. The review of the literature on a total of 85 patients with S. aureus infection revealed similar results.
Conclusions Long-acting tetracyclines may be a reasonable treatment alternative for patients with certain types of MRSA infection.
The emergence of methicillin-resistant Staphylococcus aureus (MRSA) as a cause of nosocomial and community-acquired infection has been reported worldwide [1, 2]. Intravenous vancomycin continues to be the mainstay of therapy for hospitalized patients with serious MRSA infections. However, therapy for patients with less severe infection or for those who can be treated with or switched to oral therapy is often complicated by coresistance of the organism to other classes of antibiotics. A recent study of more than 2400 MRSA isolates obtained from patients with community-onset and nosocomial infections in different regions of the United States showed coresistance to ciprofloxacin, erythromycin, and clindamycin in 89%, 93%, and 79% of all isolates, respectively [3]. In contrast, only 16% of the MRSA isolates were resistant to tetracycline, a bacteriostatic agent, making this drug class a potential treatment alternative for patients with less serious S. aureus infection who may be treated with oral therapy.
The long-acting tetracycline derivatives doxycycline and minocycline are well-absorbed by the gastrointestinal tract, have excellent tissue penetration, and demonstrate better antistaphylococcal activity than tetracycline at clinically achievable levels [4, 5]. However, clinical data on their use against MRSA infection are rare in the English literature [6]. We studied the clinical presentation and therapeutic outcome of 24 patients who received either doxycycline or minocycline for treatment of serious MRSA infection and reviewed the literature for data on the clinical efficacy of these antibiotics against S. aureus infection.
The medical records of adult patients treated with doxycycline or minocycline for tetracycline-susceptible MRSA infection at the University of Arkansas for Medical Sciences Hospital (which contains 400 beds) and the Central Arkansas Veterans Healthcare System (which contains 500 beds) were reviewed. The study period was between February 1999 and July 2004. Doxycycline or minocycline were administered orally at a dosage of 100 mg twice daily. Case patients were identified by comparing the databases of the clinical microbiology laboratory with those of the inpatient and outpatient pharmacies at both institutions. Only patients with the following types of culture-proven MRSA infection were included: complicated skin and skin-structure infection, which was defined as postsurgical wound infection, the presence of major abscess, cellulitis complicated by abscess, or the presence of infected skin ulcer [7]; urinary tract infection, which was defined as the presence of pyuria by urine analysis, the growth of ⩾105 cfu/mL by urine culture, and the appropriate clinical symptoms of infection (i.e., dysuria, frequency, urgency, or suprapubic pain), fever, and/or leukocytosis in the absence of bacteremia; and invasive infection, including bacteremia and osteomyelitis.
Patients were excluded from the study if they had only minor skin infection (folliculitis, uncomplicated cellulitis, or impetigo), if the total duration of treatment with tetracycline comprised ⩽50% of the total treatment duration that was deemed appropriate for the specific type of infection, if the follow-up time was considered inappropriate (i.e., 12 months for osteomyelitis and 3 months for all other types of infection), or if medical records were incomplete.
Patient data were obtained from the medical records using a standardized data abstraction form. Information about sociodemographic characteristics and comorbidities and previous hospitalizations and surgical procedures within the previous 6 months was obtained; comorbidity was assessed using the Charlson comorbidity scale [8]. Information about the type, antibiotic susceptibility pattern, treatment (i.e., medical and/or surgical), and outcome of each case of MRSA infection was also collected. A MRSA infection was defined as community-acquired if the organism was isolated in an outpatient clinic or within 48 h after hospital admission. A patient was defined as having a community-acquired MRSA infection without risk factors if the patient was not hospitalized within the previous 24 months; had no outpatient medical visit, nursing home admission, or antibiotic exposure within the previous 12 months; had no chronic illness; and did not use injection drugs [9]. Cure was defined as resolution of all clinical and radiological signs of infection after completion of therapy, with documentation of negative results of cultures if performed at follow-up visits and no symptoms of recurrence at the end of the follow-up period.
Susceptibilities to oxacillin, erythromycin, levofloxacin, trimethoprim-sulfamethoxazole, and tetracycline were tested by broth microdilution with the MicroScan WalkAway 96 instrument (Dade Behring), the Vitek system (BioMérieux), or the Vitek 2 system (BioMérieux), according to the NCCLS guidelines [10, 11]. We do not report on clindamycin susceptibility because the D test was not routinely performed by our hospital's laboratories on erythromycin-resistant isolates that appeared to be clindamycin-susceptible by broth microdilution testing. The study was approved by the institutional review boards of both hospitals.
We searched the English-language literature for reports of adults with S. aureus infection treated with doxycycline or minocycline, using the MEDLINE database (which includes citations from 1966 to the present). We used the search terms “Staphylococcus aureus,” “tetracycline,” “doxycycline,” “minocycline,” and “treatment.” English translations of relevant articles in the Japanese literature were not available. Studies were excluded if treatment results included only patients who were Staphylococcus carriers or had minor skin infections, as defined above; if staphylococci were not identified at the species level; if the clinical outcomes of individual patients or groups of patients with specific types of infection groups were not reported; or if adequate follow-up time was not documented. Patients infected with strains of S. aureus that were resistant to tetracycline or with polymicrobial infections were also excluded.
MRSA was isolated from 3739 clinical specimens processed by the clinical microbiology laboratories of both hospitals between February 1999 and July 2004. We identified a total of 45 patients who received treatment with the long-acting tetracyclines—doxycycline or minocycline—during this time period. Twenty-one patients were excluded because they received tetracyclines for <50% of the total treatment duration (n = 9), only minor skin infections were present (n = 6), the diagnosis of MRSA infection was not clearly established (n = 3), medical records were incomplete, or adequate follow-up was not documented (n = 3). A total of 24 patients with MRSA infection treated with doxycycline or minocycline were included in our study; 15 (63%) were from the Central Arkansas Veterans Healthcare System, and 9 (37%) were from the University of Arkansas for Medical Sciences Hospital.
Baseline demographic and clinical characteristics of these 24 patients are summarized in table 1. The median age of the cohort was 51 years, and 21 patients (88%) were male. The median Charlson comorbidity score was 2 (range, 0–5). The most common comorbid conditions in our cohort were peripheral vascular disease, in 8 patients (33%); diabetes mellitus, in 7 patients (29%); solid or hematological malignancy, in 5 patients (21%); and chronic liver disease, in 4 patients (17%). No patient was infected with HIV or had end-stage renal disease. Community-acquired MRSA was the cause of infection in 19 patients (79%). Only 1 patient (4%) with severe upper-extremity cellulitis and multiple soft-tissue abscesses after receiving a tattoo fulfilled the criteria of having community-acquired MRSA infection without risk factors. The isolate obtained from this patient was resistant to erythromycin but susceptible to levofloxacin.
Demographic and clinical characteristics of 24 adult patients with methicillin-resistant Staphylococcus aureus (MRSA) infection treated with long-acting tetracyclines.
Complicated skin and skin-structure infections were the most common clinical presentation and were seen in 16 patients (67%) (table 2). Other infections included osteomyelitis (in 4 patients), urinary tract infection (in 2 patients), septic arthritis (in 1 patient), and bacteremia (in 1 patient). MRSA was cultured from specimens obtained during surgery from 15 (71%) of 21 patients with skin and skin-structure infection, bone infection, or joint infection.
Types of infection and clinical outcomes in 24 adult patients with methicillin-resistant Staphylococcus aureus infection treated with long-acting tetracyclines.
Thirteen patients (54%) received doxycycline, and 11 patients (46%) were treated with minocycline. Minocycline was combined with rifampin in the treatment for 4 patients and combined with trimethoprim-sulfamethoxazole in the treatment for 1 patient. The median total duration of antibiotic treatment was 19 days (range, 10–72 days). Tetracyclines were administered for a mean of 89% and a median of 100% (range, 57%–100%) of the total treatment time. Eight (33%) of 24 patients had initially received intravenous vancomycin for a mean duration of 4.4 days (range, 2–6 days).
The overall clinical success rate was 83%. Treatment failed for 4 patients (17%). Minocycline-rifampin therapy failed for 1 patient with chronic osteomyelitis of the fifth metatarsal bone and severe peripheral vascular disease; this combination was also unsuccessful for a patient with septic arthritis of the shoulder and possible nonadherence. Treatment for was discontinued for 1 patient because of nausea after receiving minocycline on day 5 of treatment and for another patient because of vomiting after receiving doxycycline on day 10 of treatment. No adverse vestibular effects were reported.
We found 2 reports on the use of minocycline alone or in combination with rifampin for treatment of MRSA infection (table 3). One prospective study reported on 25 patients with severe MRSA infection, all of whom received combination therapy with minocycline and rifampin. Of those, 19 patients were cured, 1 patient with a soft-tissue infection showed improvement, and treatment failed for 5 patients [12]. Of the 5 patients for whom therapy failed, 4 had septic thrombophlebitis, osteomyelitis, liver abscess, and endocarditis (1 each). Resistance to rifampin developed in the infecting strain of MRSA in 4 patients, and resistance to minocycline emerged in the infecting strain in 1 patient. Lawlor et al. [13] reported the cure of a 45-year-old man who was an injection drug user and received 52 days of minocycline monotherapy for prosthetic valve MRSA endocarditis. Previous antibiotic regimens had failed, including almost 3 months of therapy with vancomycin and gentamicin. Another study [6] reported a successful outcome in 15 of 17 patients with staphylococcal osteomyelitis (i.e., caused by S. aureus or S. epidermidis) or soft-tissue infection. Four of these cases of infection were caused by MRSA, of which 2 were caused by a methicillin-susceptible strain. However, specific outcomes were not separately reported; therefore, we excluded the study.
Review of the literature on patients with methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-susceptible S. aureus (MSSA) infections treated with long-acting tetracyclines.
Because of the paucity of data obtained about MRSA infection, we expanded our review of the literature to include studies that reported on patients with infection due to methicillin-susceptible S. aureus [14–20]. Most of these studies were performed in the 1970s and did not specifically test for methicillin or oxacillin susceptibility; these isolates were considered to be methicillin-susceptible S. aureus. The clinical characteristics and outcomes of patients in the 7 studies included are listed in table 3.
In summary, our review of the literature revealed a total of 85 patients from 9 studies who were treated with long-acting tetracyclines: methicillin-susceptible S. aureus was the causative pathogen in 59 patients (69%), and MRSA was the causative pathogen in 26 patients (31%). Fifty-three (62%) of the patients had skin and skin-structure infection, 16 (19%) had pneumonia, and 8 (9%) had osteomyelitis. The majority of patients (81%) were treated with minocycline.
The combined studies included patients with a wide distribution of ages, different underlying diseases, and varied clinical presentations. The overall rate of cure in this heterogeneous patient cohort was 85% (72 of 85 patients). Six patients with S. aureus bacteremia, including 2 with endocarditis, were treated with minocycline; 5 of these 6 were also concurrently treated with rifampin. Therapy failed for 3 patients; 2 of them developed breakthrough infection with a rifampin-resistant strain, and 1 died from Serratia marcescens sepsis [12, 13]. Treatment with long-acting tetracyclines was administered to 8 patients with osteomyelitis (all except 1 also received concomitant treatment with rifampin). The rate of clinical failure in this subset was 38% (3 of 8 patients), whereas the microbiological failure rate was 13% (1 of 8 patients). All 3 clinical treatment failures were associated with retained infected hardware, and multiple previous antibiotic regimens had failed for 2 of the 3 patients [12, 18, 20].
Only 1 of 69 patients who received treatment with minocycline developed vestibular adverse effects manifesting as vertigo. In addition, 2 patients who received minocycline and concurrent cancer chemotherapy developed nausea [13, 14].
Oral treatment options for patients with less serious MRSA infection are limited. The majority of clinical MRSA strains from North America are susceptible to tetracyclines (∼84%), but few data exist on the efficacy of tetracyclines against infections caused by this organism.
In our case series, 24 patients had received either doxycycline or minocycline for the treatment of MRSA infection; 19 of these patients received monotherapy. Complicated skin and skin-structure infection comprised the majority of cases (67%). Twenty (83%) patients were cured by tetracycline therapy. Tetracycline therapy failed for 4 patients; treatment failure for 2 patients was determined according to clinical criteria, and 2 patients had adverse gastrointestinal effects. Of note, 1 patient with B-cell lymphoma and vancomycin allergy responded well to minocycline monotherapy for MRSA bacteremia, which likely originated from skin folliculitis. To our knowledge, only 1 other patient with MRSA bacteremia and endocarditis successfully treated with minocycline monotherapy has been described in the English-language literature [13]. Investigation of minocycline treatment in a rabbit model of MRSA endocarditis demonstrated excellent drug penetration into the cardiac vegetation and a reduction of the bacterial density in the vegetation that was comparable to that caused by vancomycin [21]. In vitro data suggest that minocycline has a better antistaphylococcal activity than does doxycycline [4, 5]. However, a clinical superiority has not been demonstrated.
Staphylococcal resistance to tetracyclines is conferred by 2 mechanisms: active reflux, which is mediated by plasmid-located tetK and tetL genes, and ribosomal protection, which is mediated by chromosomal or transposonal tetM or tetO genes [22]. Studies relating the MICs of tetracycline, doxycycline, and minocycline to the presence of these genes suggest a cross-resistance between doxycycline and tetracycline in all strains of S. aureus. In one study, 304 (76%) of 400 tetracycline-resistant MRSA strains contained the tetM gene and showed in vitro resistance to minocycline [23]. Therefore, tetracycline-resistant strains of MRSA should be considered resistant to all tetracycline class members unless testing for specific minocycline susceptibility is performed.
Of 17 isolates that were tested in our study, 16 (94%) were susceptible to trimethoprim-sulfamethoxazole. Successful use of trimethoprim-sulfamethoxazole for certain types of S. aureus infection has been previously described [24, 25]. Only 26% of MRSA strains in the United States are resistant to this drug, according to recent data from the SENTRY data [3]. However, many clinical microbiology laboratories currently use the Vitek 2 automated identification and susceptibility testing system, which does not test for staphylococcal susceptibility to trimethoprim-sulfamethoxazole [26]. Thus, clinicians might hesitate to prescribe this oral treatment alternative. Another recently available addition in our armentarium against multidrug-resistant gram-positive cocci is linezolid [27]. To date, drug resistance in S. aureus has been rarely described [28]. However, drawbacks of linezolid include its high cost, its adverse hematological effects with prolonged use, and its bacteriostatic antistaphylococcal activity.
Our literature search revealed 9 studies that reported the use of doxycycline or minocycline for methicillin-susceptible or methicillin-resistant S. aureus infection; all were published between 1971 and 1990. Additional studies on this subject in the Japanese medical literature could not be accessed [29]. The spectrum of infections, the combined cure rate (85%), and the incidence of adverse vestibular or gastrointestinal effects were similar to those in our case series. The very low rates of adverse vestibular effects in our case series and in the literature that we reviewed differ from the rates in previous reviews [30]. The reasons for this discrepancy remain unclear.
The retrospective design of our study raises the possibility of selection bias. However, the same surveillance method of comparing the databases of the clinical microbiology laboratory and the pharmacy departments was used throughout the observation period, and patients were enrolled throughout the 6-year study period. We tried to minimize misclassification and surveillance bias by using strict inclusion criteria regarding the diagnosis of MRSA infection and posttreatment follow-up. Furthermore, the heterogeneity of the patient cohorts—with respect to underlying diseases, type of infection, and antibiotic regimens in both our case series and those in the literature we reviewed—limits the generalizability of our results.
In summary, these data suggest that the long-acting tetracyclines doxycycline and minocycline might be reasonable oral treatment alternatives for patients with specific types of MRSA infection, especially those involving the skin and skin structure. Our data are not sufficient to support their use in cases of serious infection, such as bacteremia or osteomyelitis. Additional prospective data are needed to determine their efficacy and safety, compared with the efficacy and safety of other MRSA treatment alternatives.
Potential conflicts of interest. All authors: no conflicts.
IDSA Members: For your free access to this journal, log in via the IDSA members area.
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy