Results
A PubMed search using ("spinal epidural abscess" OR "spinal epidural abscesses" AND [management OR treatment]) returned a total of 429 results. Twelve articles were retrospective patient reviews comparing surgical to medical management of SEA that met our previously stated inclusion criteria (Table 1). All 12 studies were retrospective single-institution reviews; none were prospective studies of treatment. The oldest article was from 1999, and the most recently published article was from 2014. Data reported in these individual institutional reports were pooled for analysis.
Demographics
The largest case series included 355 patients, and the smallest included 30. Adding all cases together, a total of 1099 patients were included in the 12 studies. The average age range of patients was from 45 to 65 years with an overall average of 57.2 years. Sex percentages were 62.5% for males, and 37.5% for females; this was not significantly different from historical data (p = 0.13) (Table 2).
Risk Factors and Medical Comorbidities
Intravenous drug use was the most frequently reported risk factor, and it occurred in 22% of patients with SEA. This percentage was significantly higher than that reported in the historical data. Diabetes (27%) and hepatic disease (14%) were the most commonly reported medical comorbidities. Overall, risk factors and medical comorbidities were more frequently reported in the pooled current data than in the historical data (Table 3).
Pathogens
The most common pathogen found in either blood or tissue cultures was Staphylococcus aureus (63.6%), with negative cultures (13.9%), gram-negative bacteria (8.1%), coagulase-negative Staphylococcus (7.5%), and Streptococcus species (6.8%) making up much smaller percentages. In the papers reporting S. aureus methicillin sensitivity, methicillin sensitivity (38.9%) was more common than methicillin resistance (19.9%). Reihsaus et al. reported a similar breakdown of pathogens, but there were significantly more negative cultures in the current data (Table 4).
Location
The most common abscess location in the current data was the lumbar spine (48%), followed by the thoracic spine (31%) and then the cervical spine (24%). This was dissimilar to the historical data, where the thoracic spine (42.6%) was the most common location followed by lumbar (30.8%) and cervical spine segments (26.6%) (Table 5, p < 0.01).
Presenting Symptoms
In the current literature search there was a large variance of presenting symptoms that authors chose to report, with most studies reporting patients presenting with back pain or motor weakness. Percentages in the current data were taken from the total number of patients among studies reporting that symptom. The most common symptoms were back pain (67%), motor weakness (52%), fever (44%), sensory abnormalities (40%), and bladder/bowel incontinence (27%). This ordering was similar to the historical comparison, with the exception of fever being more common than weakness. There were significantly more patients with back pain, fever, and motor weakness in the historical data. In contrast, the current pooled data reported significantly more sensory abnormalities (Table 6).
Medical Versus Surgical Management
The current pooled data showed that a majority of patients with SEA received surgical treatment (60%). However, when compared with historical data, a significantly greater percentage of patients received medical management (p < 0.01), and a smaller percentage received surgical management (Fig. 2, p < 0.01). In the current pooled data, the major factor determining whether a patient received surgery was weakness or other neurological symptoms. Patient who had only back pain without other neurological sequelae were significantly more likely to receive medical management (Fig. 3, p < 0.01).
(Enlarge Image)
Figure 2.
Comparison between patients undergoing surgical and medical management for treatment of SEA with historical control. Bar graph representing distribution of surgical to nonsurgical forms of treatment from current (pooled) data versus historical control. There was a significant difference in the proportions of treatment method implemented in the current studies over the historical control (p < 0.01).
(Enlarge Image)
Figure 3.
Comparison of surgical and medical management given presenting deficits in patients with SEA. Group 1 included patients presenting without neurological deficit, with or without back pain. Group 2 included those presenting with neurological deficits, with or without back pain. The correlation between presenting deficits and management method was highly significant (Fisher's exact test, p < 0.01).
Due to the different methodologies used by studies to examine outcomes, we did not pool outcome data in our analysis; however, individual studies were analyzed for general trends (Table 7). For instance, Patel et al. used the American Spinal Injury Association (ASIA) motor score to measure outcomes differences, but other authors created unique descriptive scales. With the exception of Curry et al., none of the authors who examined statistical differences in management found a significant difference between the outcomes of patients treated medically and those treated surgically. Curry and colleagues found a statistically significant difference in medical and surgical treatment, with medical treatment failing and surgery subsequently required in a marked 49% of cases. Five authors discussed the group of patients who initially started out with medical treatment but eventually underwent surgery after their condition deteriorated, with this group accounting for 6%–49% of the cases reported in those articles. Risk factors for failure of medical management included diabetes, C-reactive protein level > 115 mg/L, leukocytosis > 12 × 10 white blood cells/L, positive blood cultures, age > 65 years, methicillin-resistant S. aureus, and advanced neurological deficit.