Epidemiology and Management of Acute Hematogenous Vertebral Osteomyelitis and Discitis in Children

Results. During the study period, 203 children were identified with AVOM with or without discitis. Ninety-two patients met the inclusion criteria (Table 1).

Table 1. Demographics, clinical presentation, microbiological etiology, diagnosis, and treatment of patients with acute vertebral osteomyelitis with/without discitis

*These organisms were identified either by culture, polymerase chain reaction (PCR) through our standardized institutional PCR, metagenomics next generation sequencing (mngs) of microbial cell-free DNA (mcfDNA) or bacterial identification through universal PCR technique or serology. **Four were readmitted unrelated to their AVOM—two for side effects related to antibiotics and one for central line related complications.

The most common clinical presentations were fever among 52 patients (56.5%) followed by back pain among 32 patients (34.8%). Twenty-nine patients (90.6%) who reported back pain were older than 5 years of age compared with only three patients younger than 5 years of age (9.4%). In children 5 years of age or younger, 34 (37%) could localize their pain compared with 58 (63%) children older than 5 years of age.

Researchers recorded the WBC, ANC, and CRP for all patients. Blood cultures were performed in 82 patients (89%). Of these, 21 were positive (25.6%). Forty-six patients (50%) had either a wound, tissue, or a bone culture done. Twenty-nine of these patients (31.5%) had no growth. The most common bacterial pathogens isolated from cultures (any cultures including blood, wound, tissue or bone culture) were S aureus (n = 29).

Of the patients with a positive tissue/bone culture, concomitant bacteremia was present in seven patients (15.5%). Four of these patients grew Methicillin Sensitive Staphylococcus aureus (MSSA) and three grew MRSA. Among the eight patients with negative blood culture, a wound/tissue, or bone biopsy yielded an etiology (MSSA = 6, MRSA = 1 and salmonella = 1). The most common empiric IV antibiotic administered was vancomycin (60.9%, n = 56). Combination therapy was used in 68 patients (73.9%) and monotherapy in 14 (15.2%). The most common combination antimicrobials were vancomycin and ceftriaxone. The total median duration of the IV and oral antibiotics administered was 62.89 days (1-554 days). The total median duration of IV antibiotics was 13.5 days (0-84 days) and total median duration of oral antibiotics was 28 days (0-539 days).

Surgical intervention in the operating room or intervention radiology occurred in 50% (n = 46) of patients. Seventeen patients (36.9%) had a bacterium isolated on cultures.  Since the most common antibiotic used in AVOM was vancomycin, we set out to determine the etiology of AVOM in patients started on empiric vancomycin. Of the 56 patients administered vancomycin, 26 (46.4%) had no growth, 17 (30.3%) grew MSSA, five (8.9%) grew MRSA, three (5.3 %) salmonella, and four (7%) grew other miscellaneous organisms. Of the 36 patients who were not on empiric vancomycin therapy, only one patient had MRSA. Among the patients who received empiric IV vancomycin therapy, most patients were switched to either oral clindamycin (n=15, 26.7%), TMP-SMX (n=16; 28.5%) or cephalexin (n=13, 23.2%) on discharge.

Comparing patients with culture positive versus culture negative AVOM. Patients with positive cultures were more likely to be older, with a mean age of 12.3 years (4.9-14.8 years) versus 7.9 years (1.7-14.2 years) (P = 0.03) in culture negative AVOM. Patients with culture positive AVOM were also more likely to have fever at presentation (n = 31; 75.6%) compared with patients with culture negative AVOM (n = 21; 41.2%; P = 0.0009). Compared with those with negative cultures, patients with culture positive AVOM had a longer length of stay (median 8 days versus 6 days; P = 0.0017) increased median duration of fever (3 days versus 0 days; P = 0.0063), and increased CRP (mean = 14.4 mg/dl versus 3.4 mg/dL; P < 0.0001). There was no difference in race/ethnicity, presenting symptoms of pain, presence of abscesses, WBC or requirement of surgical intervention. Patients with a positive culture were less likely to have discitis at presentation (n = 6; 17.6%) when compared to those with negative cultures (n = 28; 82%, P <0.0001).

Discussion. Our study shows that there is a wide variability in the management of pediatric AVOM with/without discitis. This is not surprising, as microbiological etiology is hard to obtain in children with AVOM despite the advances in diagnostic modalities. Therefore, broad-spectrum antibiotics like vancomycin continue to be the primary empiric therapy pediatric AVOM. There are few points worth noting in our study that could help clinicians formulate a management plan in these patients. A high index of suspicion is required when diagnosing AVOM, especially in children younger than 5 years of age as they are less likely to report back pain. MRI of the spine in these situations, in the absence of an alternate etiology would be helpful to prevent diagnostic delays. This is consistent with other pediatric studies, where plain radiographs were less diagnostic than MRI and delayed diagnosis of AVOM occurred, especially in the younger age group.^1,2 Second, clinicians should be aware that diagnostic yield of cultures is low in pediatric AVOM. Like other pediatric studies, where reported blood culture positivity is 20%-57%, our blood cultures were positive in only 25.6%. Tissue cultures are less commonly performed in most pediatric studies (14%-24%) and positivity rates range between 20-80%.^1,2,6,7

In our study, tissue cultures were attempted in 50% of patients and 36.9% were positive. Using molecular methods, we were able to identify patients with additional bacterium, two with Kingella kingae and one Streptococcal pneumoniae through our institutional limited 16s rRNA bacterial fluid PCR. In addition, Kingella kingae was isolated in one patient through mfDNA testing. All 10 of the expanded universal 16s rRNA bacterial identification for bacteria were negative. Given the prevalence of Kingella kingae in AVOM in preschool children,³ there could be an underestimation of its role in AVOM and discitis. In this age group, it is reasonable to include this test in diagnostic studies if a biopsy is being done. Since all the universal bacterial PCRs in our study were negative, the utility on this costly test is debatable and should be used judiciously. Given the low diagnostic yield of all the microbiological methods in AVOM and discitis, the use of multimodal diagnostic approach should be attempted in children with AVOM.  In our study, microbiological etiology was more common in children who presented with fever, older children and those with a higher CRP. Third, our study showed that although S aureus continued to be the most common pathogen, MRSA was not commonly isolated. It is interesting that while our community MRSA rates are more than 10%, we did not encounter an increase in MRSA rates with AVOM and discitis. Even among those with negative cultures, definitive therapy did not require vancomycin. Short parenteral courses on IV therapy followed by oral therapy with either clindamycin or TMP-SMX with close follow-up with an infectious diseases specialist should be reasonable for those with clinical improvement. The total duration of antibiotic therapy in our study is comparable to previous research suggesting a range of 6-12 weeks of oral and IV antibiotics.^1-8

Our study has several limitations. This is a retrospective study, so we could not determine the definitive choice of antibiotic therapy especially in a culture negative patient. Similarly, a clear recommendation for total duration of therapy cannot be made, though it seems reasonable to do a shorter course of IV therapy followed by 4-6 weeks on oral therapy in patients with AVOM who are clinically improving. Bone biopsy is an invasive procedure and given the lower yield in AVOM, we are unable to make a strong recommendation in all patients with AVOM especially if they are clinically improving. Lastly, since this is a single center study, the results may not be generalized to the rest of the US population.

Conclusion. Our study highlights the difficulty in finding an etiological diagnosis of AVOM with/without discitis in children and the variability of antibiotic management in a single institution. Short parenteral courses of IV therapy followed by oral therapy for 4-6 weeks seems to be adequate for uncomplicated cases of AVOM. The role of molecular methodology in identifying additional pathogens was not high-yield in our study, although the identification of Kingella kingae may be underestimated.