CLINICAL ARTICLE J Neurosurg Spine 26:507–512, 2017 The epidemiology of spinal tuberculosis in the United States: an analysis of 2002–2011 data *Rafael De la Garza Ramos, MD, C. Rory Goodwin, MD, PhD, Nancy Abu-Bonsrah, BS, Ali Bydon, MD, Timothy F. Witham, MD, Jean-Paul Wolinsky, MD, and Daniel M. Sciubba, MD Department of Neurosurgery, Johns Hopkins University School of Medicine, Baltimore, Maryland OBJECTIVE The aim of this study was to investigate the incidence of spinal tuberculosis (TB) in the US between 2002 and 2011. METHODS The Nationwide Inpatient Sample database from 2002 to 2011 was used to identify patients with a discharge diagnosis of TB and spinal TB. Demographic and hospital data were obtained for all admissions, and included age, sex, race, comorbid conditions, insurance status, hospital location, hospital teaching status, and hospital region. The inci- dence rate of spinal TB adjusted for population growth was calculated after application of discharge weights. RESULTS A total of 75,858 patients with a diagnosis of TB were identified, of whom 2789 had a diagnosis of spinal TB (3.7%); this represents an average of 278.9 cases per year between 2002 and 2011. The incidence of spinal TB decreased significantly—from 0.07 cases per 100,000 persons in 2002 to 0.05 cases per 100,000 in 2011 (p < 0.001), corresponding to 1 case per 2 million persons in the latter year. The median age for patients with spinal TB was 51 years, and 61% were male; 11.6% were patients with diabetes, 11.4% reported recent weight loss, and 8.1% presented with paralysis. There were 619 patients who underwent spinal surgery for TB, with the most common location being the thoracolumbar spine (61.9% of cases); 50% of patients had instrumentation of 3 or more spinal segments. CONCLUSIONS During the examined 10-year period, the incidence of spinal TB was found to significantly decrease over time in the US, reaching a rate of 1 case per 2 million persons in 2011. However, the absolute reduction was rela- tively small, suggesting that although it is uncommon, spinal TB remains a public health concern and most commonly affects male patients approximately 50 years of age. Approximately 20% of patients with spinal TB underwent surgery, most commonly in the thoracolumbar spine. https://thejns.org/doi/abs/10.3171/2016.9.SPINE16174 KEY WORDS tuberculosis; spine; Pott’s disease; epidemiology; incidence; surgery; infection PINAL tuberculosis (TB) is a known cause of spinal phosis, and paraplegia.12 Whereas spinal TB affects mostly deformity and paraplegia. First described in the children in developing nations, it is a disease of the elderly modern era by English surgeon Percival Pott in 1779 in the developed world.7 Pott’s disease develops after he- S(i.e., Pott’s disease),20 it is currently regarded as a disease matogenous spread of Mycobacterium tuberculosis from a affecting people in the developing world.7 Nevertheless, primary site (typically the lungs), which leads to destruc- TB remains a problem in developed regions such as Eu- tion of the intervertebral disc and vertebral body.11 This rope and the US.6,16,25 In 2014 alone, a total of 9421 new results in a pathological compression fracture, vertebral cases of TB were reported in the US, corresponding to an wedging, and kyphoscoliosis, which in turn may lead to incidence of 2.96 cases per 100,000 persons.5 neurological deficits due to spinal cord compression.2,14,17 Pott was the first to recognize the link between TB, ky- Despite multiple epidemiological studies examining ABBREVIATIONS ICD-9 = International Classification of Diseases, 9th Revision; LOS = length of stay; NIS = Nationwide Inpatient Sample; TB = tuberculosis. SUBMITTED February 22, 2016. ACCEPTED September 2, 2016. INCLUDE WHEN CITING Published online December 16, 2016; DOI: 10.3171/2016.9.SPINE16174. * Drs. De la Garza Ramos and Goodwin contributed equally to this work. ©AANS, 2017 J Neurosurg Spine Volume 26 • April 2017 507 Unauthenticated | Downloaded 10/01/21 11:09 PM UTC R. De la Garza Ramos et al. FIG. 1. Graph showing the population growth–adjusted incidence rate of FIG. 2. Bar graph showing the proportion of patients with spinal TB patients with a discharge diagnosis of spinal TB in the US from 2002 to among all TB cases. The proportion was found to decrease significantly, 2011. After a peak in 2003, the incidence was found to decrease signifi- from 4.1% in 2002 to 2.6% in 2011 (p < 0.001). Figure is available in cantly over time (p < 0.001). Figure is available in color online only. color online only. TB, there are limited data on the prevalence of Pott’s dis- 9,13 complication; surgical site infection; meningitis; pneumo- ease in the US. The aim of this study is to investigate nia; and incidental durotomy. Hospital charges (excluding the incidence of spinal TB in the American population be- professional fees) were also recorded and reported in 2016 tween 2002 and 2011, using a large nationwide database. US dollars. Methods Statistical Methods Study Design National estimates for spinal TB cases were calculat- This epidemiological study used the Nationwide Inpa- ed after application of discharge weights supplied by the tient Sample (NIS) database records of patients treated be- NIS. Given that this database is a 20% sample, roughly tween 2002 and 2011. Developed by the Healthcare Cost a 5-fold multiplication will give the total number of es- and Utilization Project, the NIS is the largest inpatient timated cases in the US per year. US population data for database in the US, containing hospital admission infor- each year were obtained from the website of the US Cen- mation from a 20% sample of community hospitals across sus Bureau (http://www.census.gov/popest/data/historical/ the country. This 20% sample corresponds to more than index.html). Incidences for any given year were adjusted 8 million hospital stays annually and contains diagnostic for population growth by dividing the estimate of each di- and procedure data in the form of International Classifi- agnosis (from the NIS) by the total population estimate cation of Diseases, 9th Revision (ICD-9) codes. from the US Census Bureau.4 A simple linear regression was done to analyze trends over time. Statistical analyses Patient Sample and Inclusion Criteria were performed in STATA 12 SE (StataCorp). A probabil- All patients with a diagnosis of “tuberculosis” were ity value < 0.05 was considered statistically significant. identified using the Clinical Classifications Software code #1. Patients with spinal TB were then identified via use Results of ICD-9 codes 015.00, 015.01, 015.02, 015.03, 015.04, Between 2002 and 2011, a total of 75,858 patients with 015.05, and 015.06. Patients who underwent spinal fusion a diagnosis of TB were identified, of whom 2789 had a di- were identified via ICD-9 procedural codes 81.00–81.09. agnosis of spinal TB (3.7%). The average annualized rate of spinal TB was 278.9 cases per year. During the 10-year Recorded Variables period examined, the incidence of spinal TB decreased Documented admission information included the fol- significantly, from 0.07 cases per 100,000 persons in 2002 lowing: patient age; sex; race; comorbid conditions; pri- to 0.05 per 100,000 in 2011 (Fig. 1). In 2011, this incidence mary payer (insurance); hospital region (Northeast, Mid- corresponded to 1 case per every 2 million persons. west or North Central, South, or West); hospital teaching The proportion of spinal TB cases among all TB cas- status; hospital location (urban vs rural); in-hospital mor- es was 4.1%, 5.6%, 3.3%, 3.8%, 2.6%, 3.7%, 4.0%, 4.0%, tality rate; and length of stay (LOS). For patients who un- 3.0%, and 2.6% yearly between 2002 and 2011 (Fig. derwent spinal surgery, the following complications were 2). A summary of spinal TB cases is shown in Table 1. recorded: neurological complications; iatrogenic stroke; The median age of patients was 51 years, and 61% were respiratory complications (including respiratory failure); male. There was an almost equal proportion of Caucasian pulmonary embolism; deep venous thrombosis; myocardi- (24.0%), African American (23.8%), and Hispanic pa- al infarction; acute kidney injury; urinary tract infection; tients (23.2%), and a smaller proportion of Asian/Pacific Clostridium difficile; paralytic ileus; other gastrointestinal Islanders (18.8%) and others (10.2%). The 5 most common 508 J Neurosurg Spine Volume 26 • April 2017 Unauthenticated | Downloaded 10/01/21 11:09 PM UTC Epidemiology of spinal tuberculosis in the US: 2002–2011 data TABLE 1. Summary of characteristics in 2789 patients with a TABLE 2. Summary characteristics of 619 patients who diagnosis of spinal TB underwent surgery for spinal TB between 2002 and 2011 Parameter Value Parameter Value No. of cases 2789 No. of cases 619 Median age in yrs (IQR) 51 (35–65) Median age in yrs (IQR) 52 (35–66) Age >65 yrs (%) 24.4 Age >65 yrs (%) 26.2 Sex (%) Sex (%) Female 39.0 Female 44.4 Male 61.0 Male 55.6 Race (%) Surgery location (%) Caucasian 24.0 Cervical 11.1 African American 23.8 Thoracolumbar 61.9 Hispanic 23.2 Lumbosacral 27.0 Asian or Pacific Islander 18.8 Approach (%) Other 10.2 Posterior only 38.9 Comorbidities (%) Anterior only 26.6 Alcohol abuse 3.0 Combined 34.5 Chronic lung disease 7.7 No. of instrumented levels (%) Diabetes 11.6 1–2 50.0 Drug abuse 5.5 3–7 42.0 Hypertension 28.6 ≥8 8.0 Liver disease 4.9 Complication rate (%) 38.8 Obesity 2.8 In-hospital mortality (%) 3.2 Renal failure 5.6 Median LOS in days (IQR) 16 (10–24) Paralysis 8.1 Weight loss 11.4 HIV positive 4.0 (39.9%), followed by the West (27.4%), the Northeast Malnutrition 11.6 (22.6%), and the Midwest or North Central (10.1%).