Brucellar, Pyogenic, and Tuberculous Spondylodiscitis at Tertiary Hospitals in Saudi Arabia: A Comparative Retrospective Cohort Study

Abstract Introduction Spondylodiscitis is rare yet the most common form of spinal infection. It is characterized by inflammation of the intervertebral disk space and adjacent vertebral body. In Western countries, the incidence of spondylodiscitis is increasing. Clinical outcomes most commonly reported in the literature are the 1-year mortality rate (range, 6%–12%) and neurologic deficits. Methods This multicenter retrospective cohort study assessed patients diagnosed with infectious spondylodiscitis who received treatment at King Abdulaziz Medical City in Riyadh and Jeddah, Saudi Arabia. All enrolled patients were ≥18 years old and were diagnosed per radiologic and microbiological findings and clinical manifestations between January 2017 and November 2021. Results This study enrolled 76 patients with infectious spondylodiscitis, with a median age of 61 years. All patients presented with back pain for a median 30 days. Patients were stratified into 3 groups based on the causative pathogen: brucellar spondylodiscitis (n = 52), tuberculous spondylodiscitis (n = 13), and pyogenic spondylodiscitis (n = 11). All laboratory data and biochemical markers were not significantly different. However, C-reactive protein, erythrocyte sedimentation rate, and white blood cells were significantly different in the pyogenic spondylodiscitis group, with medians of 121 mg/dL (P = .03), 82 mmol/h (P = .04), and 11.2 × 109/L (P = .014), respectively. Conclusions Back pain is a common clinical feature associated with infectious spondylodiscitis. The immense value of microbiological investigations accompanied with histologic studies in determining the causative pathogen cannot be emphasized enough. Treatment with prolonged intravenous antimicrobial therapy with surgical intervention in some cases produced a cure rate exceeding 60%.

Spondylodiscitis is rare yet the most common form of spinal infection.It is characterized by inflammation of the intervertebral disk space and adjacent vertebral body.It occurs via multiple mechanisms, including hematogenous seeding, where microorganisms travel from distant organs, and direct inoculation due to spinal surgery or trauma [1].The predisposing factors include male sex, advanced age, diabetes mellitus, cancer, penetrating trauma, and spinal surgery [2].In Western countries, the incidence of spondylodiscitis ranges from 0.4 to 2.4 per 100 000 individuals [3].
In a recent meta-analysis, 1756 cases of spondylodiscitis were analyzed: 56% of cases were male, the mean age was 60 years, and Staphylococcus spp and Mycobacterium tuberculosis were the most causative pathogens [4].The lumbosacral region was the most affected area (∼50%), possibly because of the pump effect (high vasculature), and 30% of patients developed abscesses.Neurologic compromises were observed in 27.8% of the cases [4].
The clinical outcomes most commonly reported in the literature are the 1-year mortality rate (range, 6%-12%) and neurologic deficits [5][6][7].Neurologic compromises that persist either improve or deteriorate in some patients after surgical or medical therapy, with a range of 5% to 26.7% depending on the treatment modality received, site/region of infection, association with abscess formation, and causative pathogens [5,[7][8][9][10].The goal of this study was to characterize clinical, laboratory, and microbiological findings of cases with confirmed spontaneous spondylodiscitis and to evaluate outcomes after antimicrobial therapy.

Study Setting
This multicenter retrospective cohort study assessed all patients diagnosed with infectious spondylodiscitis who received treatment at the King Abdulaziz Medical City (KAMC) in Riyadh and Jeddah, Saudi Arabia.KAMCs are accredited per the standards of Joint Commission International as 1500-and 600-bed tertiary care academic medical centers that provide care to Saudi National Guard soldiers, employees, and their families.As there was no hypothesis to be tested, we included all patients with infectious spondylodiscitis from January 2017 to November 2021 who met the inclusion criteria.Inclusion criteria were as follows: age ≥18 years; an admission diagnosis of spontaneous infectious spondylodiscitis; radiologic findings consistent with infectious spondylodiscitis per spine radiologists; clinical manifestations, including fever, back pain, and/or weight loss; microbiological findings, such as positive blood, tissue/bone, and abscess cultures or high titer in cases with brucellosis; and receipt of intravenous antimicrobial therapy for at least 6 weeks, with or without surgical intervention.Patients were excluded if they had life-threatening bacterial or fungal infections at the time of admission, required a prolonged hospital stay (>180 days), or had a penetrating spine injury or trauma.
The microbiological diagnosis of tuberculous spondylodiscitis was based on (1) the recovery of M tuberculosis from Lowenstein-Jensen agar and/or the presence of acid-fast bacilli in Ziehl-Neelsen-stained slides plus (2) caseation granulomatosis on histopathology with positive M tuberculosis polymerase chain reaction (PCR) test results from tissue, bone, or abscess specimens.The diagnosis of brucellar spondylodiscitis was based on clinical signs of brucellosis, Brucella spp growth in blood and other specimens, or serum brucella agglutination test results ≥1:320.To diagnose spontaneous spondylodiscitis caused by bacteria other than M tuberculosis or Brucella spp, bone aspiration specimens obtained through needle biopsy under computed tomography in sterile and bedside conditions were sent to the microbiology laboratory for bacterial culture.Once bacterial growth was observed in conventional cultures, automated systems and conventional methods were used for bacterial identification and antimicrobial susceptibilities.Primary outcomes include mortality at 90 and 365 days, clinical cure, and staging of neurologic compromises as compared with baseline at the end of antimicrobial therapy.Secondary outcomes were recurrence rate of spondylodiscitis within 3 months after completion of antimicrobial therapy and adverse drug reactions resulting from antimicrobial therapy.

Definitions
Clinical Cure.Clinical cure is defined as follows: • Resolution of biochemical signs of infection: normalization of white blood cells, C-reactive protein (CRP), and erythrocyte sedimentation rate (ESR) This system was used in this study during baseline assessment and diagnosis of infectious spondylodiscitis and at the end of antimicrobial therapy.

Statistical Analysis
The study participants were divided into 3 groups based on the type of spondylodiscitis: brucellar, tuberculosis, and pyogenic.Categorical data were presented as frequencies and percentages.Continuous variables were presented by the normality of their distribution-mean ± SD for normally distributed data or median (IQR) for nonnormally distributed data-overall and in groups of spondylodiscitis types.The 3 groups were compared with a χ 2 test or Fisher exact test for categorical variables.Analysis of variance and the Kruskal-Wallis test were used for continuous variables, as appropriate.The level of significance was set at P = .05.Statistical analyses were conducted with SAS version 9.4 (SAS Institute Inc).

G1: Brucellar Spondylodiscitis
Fifty-two patients were diagnosed with brucellar spondylodiscitis and were predominantly male (n = 36, 70%) with a median age of 61 years.The most common comorbidity in this group was a history of brucellosis in 42 (80.8%),followed by diabetes (46.2%) and hypertension (44.2%).Although only 20 (38.5%) had a fever, backache was consistent in all patients, occurring for 14 to 120 days before admission.Most patients evaluated for neurologic symptoms had stage 1 pain, which manifested as back pain only at the level of the affected disk space (n = 28, 53.8%), followed by stage 2, which is pain radiating from the affected spinal area (n = 8, 15.4%), and stage 3, weakness and neurologic deficit (n = 16, 30.8%); none of the patients reached stage 4, paralysis.Most patients had a history of ingesting unpasteurized dairy products and/or having animal contact.
Microbiological diagnostic investigations revealed that 29 (55.8%)patients had positive cultures: 24 and 5 had positive blood and tissue/bone cultures, respectively.The diagnosis for the rest was made with brucella titers; 23 patients had high brucella titers, with medians of 640 and 1280 for Brucella abortus and Brucella melitensis antibodies (Table 2).The lumbar spine region was the most affected for most patients (n = 41, 79%), followed by the thoracic and cervical areas.Most patients had at least 2 vertebrae infected.Further details are presented in Table 3. Eleven (21.2%) patients underwent surgical intervention, of whom 10 had spinal abscesses.Data on the types of surgical intervention and abscesses are shown in Table 3.All patients received doxycycline as the backbone of their antimicrobial therapy in combination with different agents, of which 10, 33, and 9 patients received dual, triple, and quadruple therapies.The median duration of antimicrobial therapy was 173 days (IQR, 96-255).The mortality rate was zero in the brucellar spondylodiscitis group at 90 and 365 days.The clinical cure rate was 66% (33/50) in patients followed up until the end of antimicrobial therapy (2 patients were lost to follow-up at the end of therapy).Most (80%) patients were pain-free at the end of their antimicrobial therapy; however, 9 still had back pain by the end of therapy.Three patients experienced a recurrence of brucellar spondylodiscitis within 3 months of completing treatment for the primary infection.Nineteen patients developed adverse drug reactions to therapy, including acute kidney injury (n = 2, 3.80%), gastrointestinal upset (n = 9, 17.3%), elevated liver enzymes (n = 2, 3.8%), myelosuppression (n = 2, 3.8%), and allergic reactions (n = 3, 5.8%; Table 4).

G2: Tuberculous Spondylodiscitis
Patients in G2 were younger than those in G1 or G3, with a median age of 42 years.The patients were predominantly male (69.8%), and half had a history of active tuberculosis.The duration of pain was longer, at a median 30 days, but it could extend to 240 days.The distribution of the initial staging used to evaluate neurologic deficits was as follows: stage 1 (n = 6, 46.2%), 2 (n = 2, 15.4%), 3 (n = 4, 30.8%), and 4 (n = 1, 7.7%).The diagnosis was confirmed by positive cultures and PCR; in 9 (69.2%)patients, M tuberculosis grew from tissue or bone cultures, and all patients had positive PCR results from biopsy specimens.The thoracic and lumbar areas were the most affected, and at least 2 vertebrae were infected (Table 3).Nine patients had spinal abscesses, of which 5 required surgical intervention after drainage.Most patients received first-line quadruple therapy, including rifampin, isoniazid, pyrazinamide, ethambutol, and moxifloxacin, in different combinations for M tuberculosis.The median treatment duration was 300 days (IQR, 270-455; Table 5).The primary outcomes were clinical cure 72.7% (8/11; where 2 patients were lost to follow-up appointments by the end of their treatment), no recurrence within 90 days, and no   within 90 days of completing therapy for the primary infection.Five patients developed adverse drug reactions to the therapy, including elevated liver enzyme levels (n = 2, 18.2%), acute kidney injury (n = 1, 9.1%), gastrointestinal upset (n = 1, 9.1%), and myelosuppression (n = 1, 9.1%; Table 4).

DISCUSSION
Infectious spondylodiscitis is a rare spinal infection with a low incidence [3].Recently, many studies have indicated an increased number of reported cases of infectious spondylodiscitis, up to 2.4 per 100 000 cases in Western countries [3].Local data are lacking on the incidence of infectious spondylodiscitis in patients with spinal infections.
Our study focused on evaluating all infectious spondylodiscitis cases in 2 large academic teaching centers in Saudi Arabia.Saudi Arabia is endemic for tuberculosis and brucellosis; unsurprisingly, the combinations accounted for >50% of the cases.In this study, we retrospectively followed patients with spontaneous infectious spondylodiscitis of different etiologies for 1 year; to our knowledge, this study is the first of its kind.Of the 76 patients, 52 (68.4%), 13 (17%), and 11 (14.5%) had brucellar, tuberculous, and pyogenic spondylodiscitis, respectively.Data from Turkey, southern Spain, and Italy have reported that the rate of brucellosis spondylodiscitis is between 20% and 50% [12,13].Close contact with livestock and ingesting unpasteurized dairy products, a common occurrence in some geographic areas among the Saudi population, are probable causes.
The rate of tuberculous spondylodiscitis declined at our institution when compared with the results of Al Othman et al (13 cases over 5 years vs 69 cases over 15 years) [14].In contrast, a prospective study conducted in Egypt noted that in the last 5 years, higher incidences of tuberculosis were reported among its communities (15/44) [10].Staphylococcus aureus was the most frequently isolated pathogen in patients with pyogenic spondylodiscitis, with a methicillin-resistant S aureus rate of 50%.S aureus was the most common organism associated with pyogenic spondylodiscitis, in >50% of cases with positive cultures in many previous studies [6,7,15].
The median age of patients in the current study was 61 years (IQR, 49-73), consistent with previous studies in which spondylodiscitis was reported more frequently in elderly patients [16,17].In terms of each group, patients in the tuberculous spondylodiscitis group were younger, with a median age of 42 years (IQR, 32-78), as compared with the other groups in this study and in different national and international studies, with a mean age of 50 years [14,18].Diabetes mellitus and hypertension were the most frequently observed comorbidities (41%), which is consistent with the results of other studies [7,8,15].Therefore, diabetes is considered the leading risk factor for infectious spondylodiscitis.More than 50% of our cohort had a history of active tuberculosis or brucellosis, indicating the endemicity of both diseases in the Arabian Peninsula.
Infectious spondylodiscitis is suspected when clinical symptoms are present, such as fever, localized back pain, and spinal deformity [10].Back pain was reported in all cases, whereas fever was noted in approximately 40% of patients in our study, as opposed to 54% by Waheed et al and Turunc et al [10,12].High-grade fever was more strongly associated with pyogenic spondylodiscitis (6/11 in the current study) than with tuberculous and brucellar spondylodiscitis.This finding can be  explained by the correlation between the duration of pain and the presence of fever, since 75% of the brucellar and tuberculous groups had pain for 120 and 240 days, respectively, indicating that spondylodiscitis was a late complication of tuberculosis and brucellosis.In contrast, Turunc et al found that most of their patients had spondylodiscitis early in the disease course [12].A unique evaluation of pain and accompanying neurologic manifestations was performed with Dr Darouiche's grading system and compared with other studies [11].Neurologic symptoms were most frequently reported, at approximately 50% in each group.Neurologic deficits were more common in the brucellar and tuberculous groups, with more spinal cord compression and disc/bone destruction caused by granuloma formation [12,19].Regarding spinal anatomic site involvement, the lumbar region appeared to be the most affected area, secondary to increased vasculature and proximity of infection, especially in brucellar spondylodiscitis [7,10,12,13].Most patients had at least 2 vertebral bodies involved (n = 49), 10 of whom required surgery.The decision to perform spinal surgical intervention was individualized depending on the extent of disc/bone destruction, abscess formation, and the effect on the patient's mobility.We also considered patient-specific parameters such as age and comorbidities.As surgical intervention serves as prompt source control, 17 patients underwent it with antimicrobial therapy, mostly for brucellar spondylodiscitis (n = 11).The timeline of surgical intervention ranges between the first week of hospital admission (n = 13) and 3 to 12 weeks (n = 4) after starting antimicrobial therapy.
In contrast to findings from Turunc et al and Waheed et al, inflammatory markers, including CRP and ESR, were significantly higher in the pyogenic group, which we believe is related to the disease being an early vs late complication of the primary infection [10,12].However, our results are consistent with those of Colmenero et al, where the pyogenic group had higher CRP and ESR levels and more leukocytosis [13].
Intravenous antimicrobial therapy is the cornerstone of infectious spondylodiscitis treatment, in addition to surgical intervention in some cases [3].It is worth mentioning that antimicrobials have a variable ability to penetrate bone tissue, ranging from moderate with β-lactams and glycopeptides to good or excellent with sulfamethoxazole-trimethoprim, fluoroquinolones, and rifampin [3].Combination therapy is indicated in tuberculous and brucellar spondylodiscitis secondary to the nature of the disease, which has been proven to reduce relapse episodes, lower the risk of resistance, and shorten the duration of therapy [20,21].More than 80% of patients in the brucellar spondylodiscitis group received ≥3 active agents, which may have contributed to the lower relapse rate in our cohort than in other studies [20].The brucellar spondylodiscitis group had a longer duration of therapy, at a median 173 days (IQR, 96-255), than the pyogenic group, which is consistent with the literature [17,19].In contrast to brucellar spondylodiscitis, the rules of combination therapy are not well established for pyogenic spondylodiscitis treatment [22].
Regarding the duration of antimicrobial therapy, it seems that patients with pyogenic spondylodiscitis required a shorter duration, at a median 120 days, as compared with brucellar and tuberculous spondylodiscitis (180 and 300 days, respectively).These findings are consistent with those of expert recommendations and previous studies [5,7,14].Moreover, recommendations endorsed by international guidelines for treatment of spine infections include a minimum duration of 6, 12, and up to 54 weeks for pyogenic, brucellar, and tuberculous spine infections, respectively [23,24].Indeed, 22.4% of the patients in our study underwent surgery, with a higher number for brucellar spondylodiscitis, unlike the study by Turunc et al, who reported a higher rate of tuberculous spondylodiscitis requiring surgical intervention as compared with brucellar and pyogenic spondylodiscitis [12].
Four patients had a recurrence of spondylodiscitis within 3 months of completing therapy, similar to that by Colmenero et al [13].In brucellar spondylodiscitis, 3 patients had a recurrence, 2 received dual therapy, and 1 received a short therapy duration of 70 days.The last patient with recurrence had methicillin-resistant S aureus spondylodiscitis and required dual therapy with trimethoprim-sulfamethoxazole and rifampin for 1 year.
Throughout the therapy period, side effects from antibiotics occurred in 31.6% of patients in our study, higher than that by Aagaard et al (23%) [6].The most common side effect reported by Aagaard et al was rash (n = 11) [6].In contrast, gastrointestinal upset was the predominant side effect (n = 11) in this study.
Some limitations of this study should be considered when interpreting our findings.Considering its retrospective nature, this study is at a high risk of selection and sampling biases.However, we attempted to minimize this by applying stringent inclusion criteria.The study was not sufficiently powered to detect significant differences among the groups owing to the small sample size.Finally, some patients (2 patients in G1 and G2) missed follow-up for outcome evaluation and were excluded from the final analysis.

Table 5
. Two patients died within 90 days of infectious spondylodiscitis diagnosis.The clinical cure rate was 63.6%, and 1 (9.1%) patient had a recurrence of infectious spondylodiscitis

Table 2 . Clinical and Microbiological Features of 76 Patients With Infectious Spondylodiscitis
Abbreviations: CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; PCR, polymerase chain reaction; WBC, white blood cell.a On the day when infectious spondylodiscitis diagnosis was confirmed.b Computed tomography-guided biopsy culture or surgical biopsy.c Positive PCR for M tuberculosis from biopsy specimen.

Table 4 . Outcomes of Interest
Spondylodiscitis at Tertiary Hospitals in Saudi Arabia • OFID • 5