Imipenem plus Fosfomycin as Salvage Therapy for Vertebral Osteomyelitis

CASE PRESENTATION

An 84-year-old male, with chronic kidney disease, type II diabetes mellitus, and chronic heart failure associated with mitral stenosis presented with fever, backache, and walking difficulties that lasted for 4 days. Magnetic resonance imaging (MRI) on admission showed L2-3 vertebral osteomyelitis and a bilateral psoas abscess, and the blood culture was positive for MRSA. Transthoracic echocardiography indicated no findings of vegetation; however, transesophageal echocardiography was not performed. Two weeks of VAN treatment for MRSA osteomyelitis was started, followed by 2 weeks of DAP treatment. The MIC of VAN for the MRSA isolates derived from the blood culture at this time was determined to be 1 μg/ml, whereas the MIC of DAP was not measured. The patient was transferred to Tokyo Medical University Hospital (TMUH, Tokyo, Japan) because blood cultures remained MRSA positive for 1 month and the psoas abscess deteriorated. Upon his arrival at TMUH, computed tomography (CT)-assisted drainage of the psoas abscess was performed because both blood cultures and the psoas abscess remained MRSA positive. The MICs for the isolates from the psoas abscess (TUM19798) and the blood culture (TUM19799) were determined and are shown in Table 1.

We switched to high-dose VAN treatment, combined with CT-assisted drainage, with a trough concentration of 15.0 to 20.0 μg/ml due to DAP resistance in the blood culture isolate. However, C-reactive protein (CRP) levels fluctuated, with high positive values ranging from 6.5 to 10.5 mg/dl, despite becoming negative in blood cultures on day 3 after transfer to TMUH. The treatment timeline can be summarized as 2 weeks of VAN and 2 weeks of DAP, administered at the previous hospital, followed by 4 weeks of high-dose VAN combined with abscess drainage, performed at TMUH. Despite 8 total weeks of antibiotic treatment and pus drainage, enhanced CT revealed no improvements in the remaining abscess or vertebral osteomyelitis. The CRP value was 8.4 mg/dl. Additionally, pus from the second CT-assisted drainage remained MRSA positive. Although we considered performing a more invasive surgical intervention, the patient was deemed to have a low tolerance for such intervention. We considered other antibiotic treatment regimens and the discontinuation of VAN treatment.

CHALLENGE QUESTION

What is the next option without progressing to a more invasive surgical intervention?

• A. High-dose DAP

• B. Linezolid

• C. Fosfomycin

• D. Imipenem

• E. Antibiotic combination of fosfomycin and imipenem

TREATMENT AND OUTCOME

We selected a combination treatment of imipenem/cilastatin (IPM/CS, 1.5 g per day) and fosfomycin (FOF, 4.0 g per day), based on the results of the Etest synergy test, withdrawing VAN treatment. Despite existing reports regarding the efficacy of various antibiotic combinations and because the methods for selecting ideal antibiotic combination regimens can be time-consuming, determining significant treatments can be difficult in the clinical setting. The effectiveness of each potential combination regimen depends on the response of specific MRSA isolates, which makes determining optimal antibiotic regimens difficult. The Etest synergy test has been reported to represent a practical method for assessing the efficacy of antibiotic combination regimens, without requiring the performance of time-consuming techniques (7, 8). The Etest synergy test is a prediffusion technique, in which the first or second Etest strip is removed after 1 h and replaced directly with the second or third strip (7, 8). Each Etest strip was placed so that the positions of the MICs of the drugs, which were measured a priori, overlapped.

Etest synergy tests were carried out using FOF, IPM/CS, DAP, VAN, and linezolid (LZD), in the following combinations: IPM/CS+FOF, IPM/CS+VAN, FOF+VAN, IPM/CS+DAP, FOF+DAP, IPM/CS+LZD, FOF+LZD, IPM/CS+FOF+VAN, IPM/CS+FOF+DAP, and IPM/CS+FOF+LZD. The combination effects were measured based on the distance (in millimeters) from the MIC line to the reference point. If a combination has no effect, the reference point and the MIC of each antibiotic agent will be the same, and the distance will be 0 mm. The distance of each combination regimen was evaluated to determine the effect size of each combination.

The MICs of the drugs indicated that TUM19799 was more likely to be resistant than TUM19798; therefore, an Etest synergy test was performed for TUM19799. The results of the Etest synergy test (Table 2) suggested that triple antibiotic combination regimens would be effective; however, the risks of side effects increase as the number of antibiotic agents increase. The MICs of the last antibiotic agents (in micrograms per milliliter) and the largest distances from the reference point (in millimeters) were observed for the following combinations: IPM+FOF (MIC of last antibiotic agent/distance = 0.38/8), IPM+FOF+LZD (0.38/7), IPM+FOF+VAN (0.75/6), and IPM+FOF+DAP (0.75/6); these results indicated that these combinations were more effective than any single drug and all other combination regimens.

Two weeks after the initial administration of IPM/CS and FOF, the abscess improved dramatically, and CRP values fell within the normal range, with the lowest value reported at 0.18 mg/dl. CT showed improvement of the psoas abscess, which almost disappeared. However, despite 4 weeks of successful combination therapy, the patient died of end-stage heart failure.

To elucidate the mechanisms underlying the differences in antimicrobial susceptibility between isolates, TUM19798 and TUM19799 were subjected to whole-genome sequencing using MiSeq (Illumina, Inc., CA, USA) and MinION (Oxford Nanopore Technologies, Oxford, UK), as previously described (9). Furthermore, the MiSeq reads from TUM19799 were mapped to the complete genome sequence of TUM19798, and mutations in each isolate were identified using CLC Genomics Workbench software (Qiagen). Pathogen protocols were approved by the Toho University Safety Committee for Pathogens (approval number 20-53-102).

The sequence type of each strain was ST1/SCCmec type IVa, and the spa type was t1784. The virulence genes detected were lukD, lukE, sea, seh, sek, seq, and cna. We detected several acquired resistance genes, including blaZ, mecA, ant(9)-Ia, and erm(A), and we detected single nucleotide polymorphisms (SNPs) in grlA (S80F and E84G) and gyrA (S84L). The acquired resistance genes and the SNPs resulted in resistance to β-lactams, aminoglycosides, macrolides, and quinolones. Compared to a reference S. aureus strain, MW2, two SNPs were identified in genes associated with VAN resistance, which resulted in the nonsynonymous substitutions Ile8Thr, located in prsA, and Ala8Thr, located in sigB. No SNPs were identified in genes associated with LZD or FOF resistance. The comparison between TUM19798 and TUM19799 showed highly similar overall genomic structures, with only eight distinguishing SNPs identified in the whole-genome sequences. Three detected SNPs were associated with the nonsynonymous substitutions Glu344Lys, Leu341Ser, and Asp47Gly, located in SA0802, mprF, and ald, respectively (Table 3). SNPs in mprF have previously been shown to affect the resistance of MRSA to DAP (4–6). Mutations in mprF can confer reduced susceptibility not only to exogenously administered DAP but also to endogenous cationic antimicrobial peptides; therefore, these mutations can potentially be selected without DAP selection pressure (10).

The sequencing data for TUM19798 and TUM19798 were deposited in the DNA Data Bank of Japan (DDBJ), under BioProject accession number PRJDB10142.