Background

Infection is the second leading cause of mortality, accounting for 18–25% of deaths in patients who undergo hemodialysis [1, 2]. Impairment in both cellular and humoral immunity contributes to the increased incidence of infection and infection-related hospitalization in hemodialysis patients [3,4,5]. Besides, dialysis patients have greater opportunities to receive various types of artificial medical devices for treatment and examination, catheterization for angiography and intravascular placement of catheters, and artificial grafts for dialysis access. These artificial devices greatly enhance the risk of deep organ infection including muscle abscess, frequently leading to disability and mortality in hemodialysis patients [6,7,8,9,10].

Herpes zoster is a self-limited dermatomal rash related to reactivation of the varicella zoster virus (VZV) in the dorsal root ganglion [11]. A recent population-based cohort study showed that hemodialysis patients are at 2.3 times higher risk of herpes zoster compared with the general population [12]. Standard treatment for herpes zoster and its related neuralgia includes anti-viral drugs (aciclovir, valaciclovir, famciclovir) and analgesics such as narcotics, topical lidocaine treatment, anticonvulsants, and nerve blocking [13,14,15]. Since anti-viral drugs are metabolized and secreted via the kidney, the dose should be reduced in dialysis patients. When patients with herpes zoster suffer from severe pain, epidural anesthesia is occasionally applied. However, when patients are suffering from active herpes zoster, they are immunologically more compromised than when they are not affected by that disease and put at an increased risk of infection. Here, we have a clinically important question whether epidural anesthesia is a safe treatment option for the pain control in hemodialysis patients with active herpes zoster.

We herein report the case of a patient undergoing maintenance hemodialysis who developed acute bacterial meningitis and multiple deep muscle abscesses related to epidural anesthesia during pain control for herpes zoster.

Case presentation

A 54-year-old male was hospitalized for the diagnosis and treatment for fever, backache, bilateral leg pain, and declining consciousness. He had been on maintenance hemodialysis due to endstage kidney disease with unknown etiology since he was 29 years old. At ages 46 and 49, he received surgeries for carpal tunnel syndrome secondary to dialysis-related amyloidosis. He received also total parathyroidectomy with auto-transplantation in the right forearm for the treatment of secondary hyperparathyroidism at age 47. He neither had a history of diabetes mellitus and immune deficiency nor received immunosuppressants. He had been on maintenance dialysis with the use of his arteriovenous fistula as a vascular access.

Three weeks prior to the admission, he developed herpes zoster in the left inguinal and groin area (L1–L2) and in the left waist area (L1–L2) and was treated with reduced dose of valacyclovir (250 mg/day). Because herpes zoster-related neuralgia was so severe and non-steroidal anti-inflammatory drugs were ineffective, he requested a more effective pain control and was then treated with continuous epidural anesthesia for the relief of severe pain. Seven days after starting continuous epidural anesthesia, epidural catheter was removed. At the time of catheter removal, purulent discharge was observed at the insertion site of epidural catheter. Although levofloxacin was initiated thereafter, he rapidly developed backache and bilateral leg pain and declining consciousness on the next day of catheter removal and was transferred to our hospital.

On admission, consciousness level was II-10 in the Japan Coma Scale. His height was 170 cm, his body weight was 62 kg, and his body mass index was 21.5 kg/m2. Blood pressure was 156/84 mmHg in supine position. Heart rate was 104 bpm. Body temperature was 38.4 °C. A summary of blood test is shown in Table 1 including blood hemoglobin level of 106 g/L, white blood cell count of 28,600/μL (neutrophil 89.5%), platelet count of 8.7 × 104/μL, serum albumin level of 29 g/L, and C-reactive protein level of 215 nmol/L. These results suggested mild level of anemia and malnutrition, elevated inflammatory response, and the presence of bacterial infection.

Table 1 Laboratory data on admission
Full size table

Plain computed tomography did not show any mass lesion in the head. Because bacterial meningitis was highly suspected, cerebrospinal fluid was collected by lumbar puncture. The findings of the cerebrospinal fluid showed increased cell count with higher neutrophil dominance, increased levels of protein and lactate dehydrogenase, and decreased levels of glucose, indicating bacterial infection (Table 2).

Table 2 Findings of cerebrospinal fluid on admission
Full size table

Magnetic resonance imaging was performed for further evaluation. In the T2-weighted imaging, round high-intensity lesion at the L4/L5 level, indicative of epidural abscess, was observed (Fig. 1a). Also, round high-intensity lesions in the left iliopsoas muscle and in the bilateral erector spinae muscles were detected (Fig. 1b). Finally, he was diagnosed with bacterial meningitis and multiple abscesses in the iliopsoas and erector spinae muscles, which were probably triggered by the indwelling catheter for continuous epidural anesthesia.

Fig. 1
figure 1

Magnetic resonance imaging of the sagittal and cross-sectional spine. a Epidural abscess (yellow arrow) shown as a rounded high-intensity lesion at the L4/L5 level. b Abscess in the left iliopsoas muscle (yellow arrow, round high-intensity lesion) and bilateral erector spinae muscles (red arrows, round high-intensity lesions). L lumbar

Full size image

Treatment with immunoglobulin (5 g/day for 3 days), antibiotics (vancomycin, 2 g/week; ampicillin 1 g/day; ceftriaxone, 2 g/day), and phenytoin were initiated on the first hospital day. On the fifth hospital day, blood culture turned out to be negative. However, culture result of cerebrospinal fluid disclosed Staphylococcus aureus. The Staphylococcus aureus was sensitive to cefazolin, gentamycin, amikacin, erythromycin, clindamycin, minocycline, vancomycin, teikoplanin, levofloxacin, and sulfamethoxazole/trimethoprim but resistant to ampicillin. Accordingly, ampicillin was discontinued and clindamycin was added instead. However, high fever lasted and serum inflammatory markers did not decrease. Hence, minocycline and sulfamethoxazole were added on after consulting our infection team. On the 14th hospital day, cell count and lactated dehydrogenase level in the cerebrospinal fluid decreased compared to the initial evaluation. His consciousness returned to almost normal. However, fever lasted and serum levels of C-reactive protein remained high. At this point, echocardiography did not show any evidence of infective endocarditis, but gallium scintigraphy showed a strong accumulation around the left hip joint. Magnetic resonance imaging performed on the 19th hospital day showed bursitis around the femoral head and pyomyositis in the left gluteus minimus and medius muscles, large abscess in the left vastus lateralis muscle, which was probably spread from the abscess in the left iliopsoas muscle (Fig. 2). Although intravenous antibiotic administration (meropenem 1 g/day and vancomycin 2 g/week) was continued, multiple abscesses did not shrink. In addition, repeated culture of blood and cerebrospinal fluid did not detect any bacteria. Accordingly, surgical open drainage was performed on the 48th hospital day, followed by a drastic shrink of abscess in the vastus lateralis muscle and a decrease in serum inflammatory markers. Meropenem (1 g/day) was continued even after abscess disappeared. During the hospitalization, he also developed disuse atrophy and needed rehabilitation because of the long-term bedridden condition. On the 108th hospital day, he was finally transferred to another hospital for rehabilitation and maintenance hemodialysis. At 1 year after his discharge, he had completely recovered from multiple deep muscle abscesses and was on maintenance hemodialysis without relapse. Clinical course during the hospitalization is summarized in Fig. 3.

Fig. 2
figure 2

Magnetic resonance imaging of the abdomen. a Bursitis around hip joints (yellow arrows) and pyomyositis in the left gluteus minimus (red arrow) and gluteus medius muscle (white arrow heads). b Abscess in the left gluteus medius (yellow arrow) and vastus lateralis muscle (red arrow) (cross-section). c Abscess in the left gluteus minimus and medius muscles (yellow arrow) and vastus lateralis muscle (red arrow) (coronal section)

Full size image
Fig. 3
figure 3

Clinical course after hospitalization. ABPC ampicillin, CLDM clindamycin, CSF cerebrospinal fluid, CRP C-reactive protein, CTRX ceftriaxone, IMP/CS imipenem/cilastatin, MEPM meropenem, MINO minocycline, MRI magnetic resonance imaging, MSSA methicillin-sensitive Staphylococcus aureus SMX sulfamethoxazole, TMP trimethoprim, VCM vancomycin, WBC white blood cell

Full size image

Discussion

Herpes zoster is a process of reactivation of VZV which remains dormant in the sensory ganglia of the cranial nerve or the dorsal root ganglia after a previous varicella infection [16, 17]. Epidural anesthesia, occasionally used for pain control caused by various disorders in the general population, is also effective in hemodialysis patients. However, when patients are being affected by herpes zoster, it is unclear whether epidural anesthesia for pain control is a safe treatment, especially in hemodialysis patients. Generally, patients with active herpes zoster are immunologically being more compromised than when they are not affected by reactivation of VZV. As for hemodialysis patients, they are immunologically compromised because of their impaired immunity by uremic milieu. Besides, artificial devices such as catheter increase the risk of infection. Considering these medical conditions, epidural anesthesia may not be a good therapeutic option for the control of herpes zoster-related pain in hemodialysis patients with active herpes zoster, potentially causing severe infection as in the present case. Hence, further studies are necessary to determine whether the risk for epidural catheter-related infection in hemodialysis patients with active herpes zoster is higher than those without active herpes zoster.

Psoas muscle abscess is a rare but lethal complication that is occasionally seen in hemodialysis patients [8,9,10, 18,19,20,21,22,23,24,25,26,27,28]. Triad of the classical psoas muscle abscess consists of fever, pain, and limp [29]. Psoas muscle abscess is generally classified into primary and secondary form, depending on the presence or absence of the underlying disease. Primary psoas abscess is caused by bloodstream-mediated infection from a distant site by bacteremia or sepsis. Secondary psoas abscess is usually related to direct spreading or expansion from the adjacent organs and often caused by inflammation and neoplasm of the intrapelvic organs such as inflammatory bowel diseases [23]. Risk factors of psoas muscle abscess include diabetes mellitus, intravenous drug use, immunosuppressive drugs, and human-immunodeficiency-virus infection [29]. Diagnosis often requires imaging such as computed tomography and magnetic resonance imaging and occasionally positron emission tomography. As for the causative bacteria, Staphylococcus aureus is the major pathogen detected in those reported cases. Notably, increasing number of cases is now caused by methicillin-resistant Staphylococcus aureus. Treatment involves intravenous administration of sensitive antibiotic agents for at least several weeks and interventions such as needle aspiration or open surgical drainage and debridement of the affected muscles [21]. A previous report showed that the prognosis of secondary psoas abscess is worse than primary psoas abscess, and unless open surgical drainage is performed, its prognosis is unacceptable [30].

Table 3 summarized the previous 22 reports including our case on psoas muscle abscess in hemodialysis patients [8,9,10, 19,20,21,22,23,24,25,26,27,28]. The age of the patients ranged from 45 to 78 years. Both male and female patients were affected. Primary infection sites included catheters for vascular access, enterocolitis, and discitis. The causative pathogens included Staphylococcus aureus and enteric bacteria, depending on the primary source of infection. All the catheter-related psoas muscle abscesses were caused by methicillin-sensitive or resistant Staphylococcus aureus. Most cases are treated with combination of intravenous antibiotics treatment and surgical drainage, leading to remission of psoas muscle abscess. As for psoas muscle abscess secondary to catheter-related infection, all previous cases were related to catheter used for vascular access and catheters inserted into the jugular, subclavian, or femoral veins for temporal or permanent use [7, 8, 19, 20, 22, 25, 27]. In this regard, our case is the first to report psoas muscle abscess caused by epidural catheter used for continuous epidural anesthesia.

Table 3 Review of the literature on psoas muscle abscess in patients undergoing hemodialysis
Full size table

Bacterial infection in the erector spinae muscle is not so common. Staphylococcus aureus occasionally causes abscess in the erector spinae muscle, named tropical myositis, which is restricted in the tropical regions [31, 32]. However, this disease entity is somewhat different from erector spinae muscle abscess caused by bacteremia. To our knowledge, abscess in the erector spinae muscle has not been reported in dialysis population. In the present case, Staphylococcus aureus was the causative pathogen and successfully treated by sensitive antibiotics. Our case indicates that abscess in the erector spinae muscle can be the possible cause of back pain in a particular medical condition. As for the route of abscess in the erector spinae muscle, two potential pathways were presumed. One was via bloodstream-mediated dissemination and should be prevented by controlling the primary infection promptly. Another was via the route made by lumbar punctures for continuous epidural anesthesia. Because lumbar puncture through para-median pathway penetrates erector spinae muscles, multiple punctures could cause infection of that muscle. In this regard, it may be said that epidural anesthesia may sometimes cause spinae muscle abscess especially in hemodialysis patients because epidural puncture is relatively difficult to do on dialysis patients and sometimes needs multiple punctures via the para-median pathway.

Abscess in the vastus lateralis muscles caused by non-tuberculous bacteria has rarely been reported both in the general population and hemodialysis patients. Since muscle abscess without trauma is usually caused by contamination of the bacteria into the bloodstream, it is reasonable to think that the site of abscess may be stochastic and there may be no common site of muscle abscess. Or blood supply to muscle may be more diverse depending on the location of the muscle, since the psoas muscle is known to have a huge blood supply compared to other muscles. This may be one of the reasons why cases of psoas muscle abscess have been the most. The mechanisms underlying bacterial lodging to muscle may be more complicated. Hence, further studies are necessary to determine why some muscles are prone to develop abscess and others are not under bacteremia or sepsis.

Drainage at an appropriate timing is critical in the management of deep muscle abscess including psoas muscle abscess. Open surgical drainage and percutaneous drainage are the two common modalities for abscess drainage. However, a previous study showed that open surgical drainage reduced total hospital days compared with needle aspiration, especially when abscess is larger [33]. As shown in Table 3, most of the remission cases experienced drainage as well as administration of sensitive antibiotics. In our case, open surgical drainage was performed at 29 days after psoas muscle abscess was detected by imaging. In our case, delayed surgical drainage might be responsible for the delayed remission of our case. One of the important messages of our case was that open surgical drainage without delay is necessary when abscess is larger and administration of sensitive antibiotics for 5–7 days is ineffective [26]. The importance of open surgical drainage is probably explained by the fact that delivery rate of antibiotic to the abscess site is lower than that of other infection sites [34].

As for the route of infection, primary deep muscle infection including iliopsoas muscle abscess is usually mediated by the bloodstream, namely secondary to bacteremia and/or sepsis [10, 18, 28]. However, judging from the distribution and progression of the multiple abscesses in the left-sided muscles, we cannot deny the possibility that abscess in the left iliopsoas muscle expanded into abscess in the left gluteus muscle and vastus lateralis muscle. As for dialysis population, vascular access is often the primary source of infection, even when vascular access infection looks absent [35]. In our case, considering that the epidural space is not directly connected to the iliopsoas muscle and erector spinae muscles and the muscle abscesses distributed bilaterally in our case, it is plausible to think that dissemination of Staphylococcus aureus in the epidural space via the bloodstream was ascribed to the multiple deep muscle abscesses. These results indicate that early control of bacterial meningitis with appropriate antibiotics is critically important to prevent secondary bacteremia and distant organ infection such as muscle abscess that can lead to life-threatening conditions.

Finally, it would be interesting and educational to discuss the best approach to this case. When purulent discharge from the insertion site of epidural catheter is noted, the catheter should be removed and purulent discharge and blood should be cultured to detect the causative organism. Repeat blood cultures until the causative pathogen is detected. Until the causative bacteria are identified, empiric therapy should be started with broad-spectrum antibiotics covering Staphylococcus aureus and Gram-positive bacteria, since these bacteria are the two major pathogens for epidural catheter-related infection. According to the guideline for bacterial meningitis in Japan, carbapenem and third-generation cephalosporins such as ceftriaxone are recommended because these drugs show a good penetration to the central nervous system [36, 37]. In recent years, the incidence of psoas abscess caused by methicillin-resistant Staphylococcus aureus is increasing. Hence, vancomycin should be included. Once the causative bacteria is  detected and its susceptibility to antibiotics is determined, unneeded antibiotics should be discontinued and sensitive antibiotics should be continued for at least 4–6 weeks. If patients complicate bacteremia/septicemia/sepsis and distant organ infection is suspected as in the present case; imaging such as computed tomography and magnetic resonance imaging should be immediately performed to identify the location of the distant organ infection, then consider if there is a need for surgical intervention. Echocardiography should also be performed since dialysis population often are at increased risk for infective endocarditis. When the presence of intravascular infection or bloodstream-associated infection is suspected, continue treatment with sensitive antibiotics for weeks after blood culture becomes negative.

Conclusions

We presented here the case of a male patient receiving maintenance hemodialysis, whose bacterial meningitis and multiple abscess in the deep muscles were related to the indwelling epidural catheter and were successfully treated with antibiotics treatment and open surgical drainage. Our case confirms that treatment with sensitive antibiotics and open surgical drainage at an appropriate time is necessary to secondary deep muscle abscess in hemodialysis patients. Furthermore, our case suggests that epidural anesthesia may not be a good therapeutic option for the control of herpes zoster-related pain because patients with active herpes zoster is immunologically more compromised during and shortly after reactivation of VZV than when they are not affected by herpes zoster, especially in hemodialysis patients, who are already immunologically more compromised than the general population.