|Year : 2020 | Volume
| Issue : 2 | Page : 167-172
Hospitalizations for tuberculous peritonitis in the United States: Results from the national inpatient sample database from 2002 to 2014
Eric Martin Sieloff1, Adam Timothy Ladzinski1, Neiberg de Alcantara Lima1, Duncan Vos2, Harry Boamah3, Thomas Andrew Melgar4
1 Department of Internal Medicine, Adolescent and Internal Medicine, Western Michigan University Homer Stryker M D School of Medicine, Kalamazoo, MI, USA
2 Department of Biostatistics, Adolescent and Internal Medicine, Western Michigan University Homer Stryker M D School of Medicine, Kalamazoo, MI, USA
3 Department of Infectious Diseases, Adolescent and Internal Medicine, Western Michigan University Homer Stryker M D School of Medicine, Kalamazoo, MI, USA
4 Department of Pediatrics, Adolescent and Internal Medicine, Western Michigan University Homer Stryker M D School of Medicine, Kalamazoo, MI, USA
|Date of Web Publication||29-May-2020|
Eric Martin Sieloff
Department of Internal Medicine, Western Michigan University Homer Stryker M D School of Medicine, Kalamazoo, MI
Source of Support: None, Conflict of Interest: None
Background: Tuberculosis (TB) is an uncommon disease in the Western hemisphere that can present with peritoneal involvement, as tuberculous peritonitis(TBP) causing abdominal pain and fever. The healthcare and economic burden of TBP in the United States remains unknown. Methods: The National Inpatient Sample database was utilized to investigate TBP hospitalizations from 2002 to 2014. Economic expenditures, patient and hospital demographics, and associations of certain comorbidities with TBP were analyzed. Results: A total of 5878 hospitalizations for TBP occurred over the 12-year duration, with $420 million in-hospital charges. The median patient age was 45 years (interquartile range: 31.1–61.7), with the majority being Hispanic (27.15%). Hospitalizations occurred primarily in the Western (31.3%) and Southern (31.7%) United States. Patient comorbidities and the respective odds ratio associated with TBP included HIV (33.56), continuous peritoneal dialysis (10.49), malnutrition (7.38), liver cirrhosis (6.87), and liver cirrhosis sequelae (6.91). Nearly 6.37% of TBP hospitalizations also had active pulmonary TB. Conclusion: Although TBP is uncommon in the United States, it should be considered in patients presenting with abdominal pain and fever and a history of HIV, continuous peritoneal dialysis, malnutrition, liver cirrhosis, or liver cirrhosis sequelae.
Keywords: Epidemiology, peritonitis, risk factors, tuberculosis, United States
|How to cite this article:|
Sieloff EM, Ladzinski AT, Alcantara Lima Nd, Vos D, Boamah H, Melgar TA. Hospitalizations for tuberculous peritonitis in the United States: Results from the national inpatient sample database from 2002 to 2014. Int J Mycobacteriol 2020;9:167-72
|How to cite this URL:|
Sieloff EM, Ladzinski AT, Alcantara Lima Nd, Vos D, Boamah H, Melgar TA. Hospitalizations for tuberculous peritonitis in the United States: Results from the national inpatient sample database from 2002 to 2014. Int J Mycobacteriol [serial online] 2020 [cited 2020 Jul 11];9:167-72. Available from: http://www.ijmyco.org/text.asp?2020/9/2/167/285238
| Introduction|| |
Tuberculosis (TB) is a prominent disease worldwide, particularly in developing countries. While any organ system can be infected by TB, pulmonary TB is the most prevalent; yet prior to the development of anti-TB drugs, approximately 50%–90% of patients also had gastrointestinal manifestations. With modern medicine, this incidence is now less prominent but remains significant.
Tuberculous peritonitis (TBP) is an extrapulmonary manifestation of TB that involves Mycobacterium seeding of the peritoneum. It is estimated that the prevalence of TBP has decreased and now comprises 4%–10% of all extrapulmonary cases of TB worldwide., Despite its scarcity, the American HIV epidemic caused a resurgence in TB cases including TBP. If undiagnosed and untreated, the mortality of TBP can be up to 50%–70%.
TBP often presents with symptoms of ascites, abdominal pain, night sweats, and fever., The gold standard for TBP diagnosis is either an ascitic fluid culture of Mycobacterium (which can often take weeks to result) or a surgical laparoscopy with peritoneal biopsies. Elevated adenosine deaminase levels can suggest a diagnosis of TB with a lesser sensitivity and specificity of 94% and 92%, respectively.,
Abdominal pain and fever are patient complaints frequently faced by physicians, yet current literature does not clearly indicate when the differential diagnosis of TBP should be considered in the developed countries of the western hemisphere. Established risk factors for TBP in developing countries of the Eastern hemisphere include HIV, liver cirrhosis, renal failure with continuous ambulatory peritoneal dialysis (CAPD), immunosuppression, and malignancy; however, patient demographics vary substantially between the Eastern and Western hemispheres, further separated when comparing developed countries with developing countries.,,, This study was performed to characterize the current in-hospital prevalence and affected patient demographics of TBP in the United States and to determine the co-occurring comorbidities that could be considered risk factors for TBP.
| Methods|| |
The National Inpatient Sample (NIS) database was used to assess the frequency and trend of in-hospital TBP in the United States from 2002 to 2014. It contains hospital discharge information from the State Inpatient Databases and is designed to represent hospital discharges at a national level using stratified random sampling stratified by the US census region, urban or rural location, teaching status, ownership, and bed size.
Descriptive analysis was used to assess the prevalence of in-hospital TBP and patient characteristics, as indicated by the inclusion criteria's diagnosis codes for the study period. Inclusion criteria consisted of the presence of at least one indication for TBP by International Classification of Diseases (ICD)-9 codes 14.00–14.06 and 14.80–14.86.
Categorical patient characteristics are reported as frequencies (percent) and continuous patient characteristics are reported as median values (interquartile range). To assess the frequency of co-occurrences with TBP, ICD-9 codes were used to create composite variables for HIV, Type 1 diabetes, Type 2 diabetes, immunosuppression, obesity, morbid obesity, malnutrition, organ transplant recipient, chronic kidney disease (CKD), sickle cell anemia, peripheral vascular disease, rheumatoid arthritis, alcohol abuse, malignancy, hematologic malignancy, incision and drainage, cirrhosis of the liver, liver cirrhosis sequelae, chronic hepatitis, alcohol abuse and at-risk alcohol use, hepatitis B and hepatitis C. In addition, ICD-9 procedural codes were used to identify TBP discharges associated with the following procedures: CAPD, exploratory laparotomy, laparoscopy, biopsy of the peritoneum, percutaneous biopsy of an intra-abdominal mass, peritoneal lavage, and percutaneous drainage.
The prevalence of co-occurring extraperitoneal TB in addition to TBP was analyzed including active pulmonary TB and nonpulmonary extraperitoneal TB as indicated by the respective ICD-9 codes.
Weighted frequencies were reported, and all analyses were completed using weighted estimates in accordance with the NIS sampling methodology. SAS Studio 9.4 (SAS Institute, Inc., NC, USA) was utilized for analysis. Data collected from the NIS database are publicly available de-identified patient information and therefore, review and approval of the Institutional Review Board at Western Michigan University was not required.
| Results|| |
Between the years 2002 and 2014, the NIS database yielded a weighted total of 5878 TBP discharges for which the greatest number occurred in 2002 and the fewest in 2011 [Figure 1].
|Figure 1: Tuberculous peritonitis hospitalizations by year relative to all hospitalizations|
Click here to view
Patients' age, hospital length of stay, gender distribution, ethnicity, quartile of income, and insurance payer data are summarized in [Table 1]. There were 230 (3.91%) in-hospital deaths associated with TBP throughout the study period. The total hospitalization charges associated with TBP was $420,199,765 from 2002 to 2014, with annual charges shown in [Table 2].
The majority of TBP cases were seen at urban hospitals: nearly 29.24% of discharges occurred at urban teaching hospitals and 66.62% of discharges occurred at urban nonteaching hospitals with 4.15% of discharges from rural hospitals – which can be compared to hospital characteristics for non-TBP patients analyzed throughout the same time period: 45.54% of patients were at urban teaching hospitals, 41.59% were at urban nonteaching hospitals, and 12.87% at rural hospitals. Further, TBP hospitalization geographical data are shown in [Table 1] [Figure 2].
|Figure 2: Census regions and divisions of the United States per National Inpatient Sample analysis|
Click here to view
The procedures associated with TBP hospitalizations are shown in [Table 3]. Percutaneous abdominal drainage was the most frequent procedure performed, for 1314 hospitalizations.
|Table 3: Procedures frequency performed in association with tuberculous peritonitis hospitalizations|
Click here to view
The greatest number of co-occurrences coded at discharge with TB peritonitis included HIV, CKD, alcohol abuse, and liver cirrhosis; HIV and CAPD had the greatest odds ratios (ORs) associated with TBP at 33.56 and 10.54, respectively [Table 4]. There were no diagnoses of immunosuppression or incision and drainage coded with the diagnosis of TBP. There were very few cases (≤10, too few to report per the NIS agreement) of TBP occurrences with sickle cell anemia, morbid obesity, and chronic hepatitis, hepatitis C, and hepatitis B. There was a statistically insignificant co-occurrence rate of peripheral vascular disease, Type 1 diabetes, and obesity associated with TB peritonitis. There were no discharges of TB peritonitis associated with an ICD-9 code for imprisonment.
|Table 4: Co-occurrence of tuberculous peritonitis with comorbid illnesses|
Click here to view
A total of 375 TBP discharges (6.37%) were associated with pulmonary TB; conversely throughout the same study period, 0.73% of pulmonary TB discharges also had TBP. Patients in the latter group had an OR of 19.23 (95% confidence interval [CI]: 12.05–30.69) for having HIV, 3.08 (95% CI: 1.40–6.75) for having alcohol abuse, and 0.61 (95% CI: 0.279–1.34) for having CKD. Further analysis yielded that 76 TBP discharges (1.29%) had co-occurring extrapulmonary nonperitoneal TB – yet, due to low frequency, further comorbidity association analysis could not be performed.
| Discussion|| |
This is the first epidemiological database study to assess the prevalence and co-occurrence of comorbid factors for in-hospital TBP within the United States. Between the years 2002 and 2014, there were nearly 6000 hospitalizations for TBP for which the median hospitalization charge was $40,602. Figures shown in [Table 1] can be compared to recent values published for patients with spontaneous bacterial peritonitis (SBP) – which can present with similar symptoms: Caucasians (62.4%) made up the largest racial group and men (63.9%) were more likely to be hospitalized with SBP than women. Medicare insurance was the most common payer (37.3%) and hospital charges range from $73,533 to $138,273. SBP is far more prevalent than TBP, with 140,430 hospitalizations recorded between 2006 and 2014.
TBP hospitalizations were more prevalent at urban nonteaching hospitals compared to the general hospitalization between 2002 and 2014. Human immunodeficiency virus had the greatest ORs: those with HIV were found to have an OR for having TBP of 33.56. With HIV, individuals have impaired Th1 immune response, making them susceptible to Mycobacterium tuberculosis. This immunosuppression is likely what influenced a high OR of HIV associated with patients with pulmonary TB and TBP. Previous studies have demonstrated this HIV association with TBP and HIV in developing countries.,,,
CAPD had the second greatest OR with TBP at 10.49, suggesting a strong association with developing TBP in the United States, as in the Eastern hemisphere. The relative prevalence of patients on CAPD in the United States was 0.01% of the population in 2011, similar to figures reported in Taiwan in 2009 (0.025% of population) and Turkey in 2008 (0.01%) where previous studies have demonstrated CAPD as being a risk factor for TBP.,, CKD has previously been demonstrated to have a strong association with TB including TBP and other uncommon presentations such as tuberculous pericarditis., End-stage renal disease and CKD are associated with uremia, resulting in diminished innate immune defenses against TB, thus predisposing such individuals to TBP.
Alcohol abuse can often be associated with malnutrition in both compensated and decompensated liver cirrhosis, for which all factors were found to be associated with TBP in the United States – as has previously been demonstrated., There was no significant association of hepatitis B or C or obesity with TBP, suggesting alcohol as being the sole cause for cirrhosis in cases associated with TBP.
With general malignancy having been a previously suspected risk factor for TBP, our results demonstrate that the association holds only for malignancy of hematologic origin in the United States., These data show that there is no apparent association between TBP and “malignancy” in general, suggesting that cases of solid tumor malignancies may not be associated with TBP. However, due to NIS data being discharge specific rather than patient specific, this population may be misrepresented due to the frequent need of hospitalizations for treatment of both hematological and nonhematological malignancies for disease-specific treatments. Though the disease association with TBP may be skewed, the increased economic burden of hematological malignancies associated with TBP remains significant.
Geographical analysis of TBP within the United States showed that most cases occurred in the Southern and Western United States. Immigration likely played a role in this distribution as individuals who have immigrated from South America or the Middle East – both regions with a higher TB prevalence – settled primarily in the Western and Southern states as suggested by the immigrant geographical settlement data from the Migration Policy Institute., The NIS demographical data cannot be further specified by region – only by subdivisions as listed in [Figure 2].
Abdominal procedures – both surgical and percutaneous – were reportedly performed in 1/3rd of the patients discharged with TBP. The gold standard of diagnosis for TBP includes a time-consuming ascitic fluid Mycobacterium culture or more immediate intraperitoneal biopsy. Biopsies were collected for only 17.72% of hospitalizations, though the 22.36% of hospitalizations with paracentesis performed may have been diagnosed by Mycobacterium ascites culture, or by using lesser accurate ascitic fluid analysis – though less accurate, the former is able to differentiate the diagnosis from similarly appearing pathologies on imaging which include peritoneal carcinomatosis, sarcoidosis, and Crohn's disease., Ultimately, the method of diagnosis for TBP was likely based on physician preference. Another explanation for this low procedural frequency could be that previously diagnosed TBP patients were hospitalized multiple times during their treatment course.
The 6.37% of TBP hospitalizations associated with pulmonary TB, the 0.73% of pulmonary TB hospitalizations associated with TBP, and the 1.29% TBP hospitalizations associated with extrapulmonary nonperitoneal TB are significantly different figures from the Eastern hemisphere, where in Taiwan, 68.2% of TBP patients were found to have co-occurring extraperitoneal TB, for which 36% of cases were active pulmonary TB. This suggests that in the United States, TBP is less commonly the result of a co-occurring active Mycobacterium infection located elsewhere.
We acknowledge several limitations within this study. With this retrospective database analysis inherently being driven by diagnosis codes, all associated biases exist including those due to human error. Regardless, diagnostic codes have frequently been used in the field of medical research to establish valid associations and trends, for which the NIS is the largest database to provide such information. However, as the NIS database is discharge specific and not patient specific, it is not possible to distinguish repeat admissions by individual patients. Further, only in-hospital outcomes can be investigated, and not long-term outcomes. Certain comorbidities may be overrepresented within the hospital compared to others (particularly hematological malignancies and general malignancies with associated hospitalizations for chemotherapy, along with diabetes or CKD) that may skew their co-occurring ORs with TBP and thus they may be even greater risk factors for TBP than presented herein. However, the substantial use of the NIS databases suggests that significant resultant trends and associations still bear merit. Even though specific associations may be skewed, the overall economic burden (hospital charges) associated with these diseases remains significant. Further, another significant limitation is that with small sample sizes (<11 unweighted associations), the NIS database user license agreement cannot provide specific quantitative figures in an effort to protect patient privacy. Future research will be needed to determine the long-term prognosis of such patients with TBP after discharge in the United States. Regardless of these limitations, the NIS data provided herein provide valuable information establishing the patient demographics, relative associations, and the national economic burden associated with TBP.
| Conclusions|| |
TB is an uncommon cause of peritoneal infection, yet such cases do continue to occur in the United States causing significant morbidity, mortality, and economic burden. When a physician is faced with patient complaints of abdominal pain and fever in the United States, with the evidence presented herein, we suggest that TBP be considered in the differential diagnosis for patients with a history of HIV, CKD (particularly those using CAPD), malnutrition, alcohol abuse, both decompensated and compensated cirrhosis, organ transplant, rheumatoid arthritis, and hematological malignancy. With an increased understanding of TBP, physicians may be able to recognize this disease rapidly and initiate treatment judiciously in an attempt to reduce long-term patient mortality and the health-care economic burden associated with TBP.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Chou CH, Ho MW, Ho CM, Lin PC, Weng CY, Chen TC, et al
. Abdominal tuberculosis in adult: 10-year experience in a teaching hospital in central Taiwan. J Microbiol Immunol Infect 2010;43:395-400.
Sochocky S. Tuberculous peritonitis. A review of 100 cases. Am Rev Respir Dis 1967;95:398-401.
Demir K, Okten A, Kaymakoglu S, Dincer D, Besisik F, Cevikbas U, et al
. Tuberculous peritonitis – Reports of 26 cases, detailing diagnostic and therapeutic problems. Eur J Gastroenterol Hepatol 2001;13:581-5.
Mehta JB, Dutt A, Harvill L, Mathews KM. Epidemiology of extrapulmonary tuberculosis. A comparative analysis with pre-AIDS era. Chest 1991;99:1134-8.
Chow KM, Chow VC, Hung LC, Wong SM, Szeto CC. Tuberculous peritonitis-associated mortality is high among patients waiting for the results of mycobacterial cultures of ascitic fluid samples. Clin Infect Dis 2002;35:409-13.
Guirat A, Koubaa M, Mzali R, Abid B, Ellouz S, Affes N, et al
. Peritoneal tuberculosis. Clin Res Hepatol Gastroenterol 2011;35:60-9.
Shakil AO, Korula J, Kanel GC, Murray NG, Reynolds TB. Diagnostic features of tuberculous peritonitis in the absence and presence of chronic liver disease: A case control study. Am J Med 1996;100:179-85.
Burgess LJ, Swanepoel CG, Taljaard JJ. The use of adenosine deaminase as a diagnostic tool for peritoneal tuberculosis. Tuberculosis (Edinb) 2001;81:243-8.
Sanai FM, Bzeizi KI. Systematic review: Tuberculous peritonitis – Presenting features, diagnostic strategies and treatment. Aliment Pharmacol Ther 2005;22:685-700.
Suntur BM, Kuşçu F. Pooled analysis of 163 published tuberculous peritonitis cases from Turkey. Turk J Med Sci 2018;48:311-7.
Yeh HF, Chiu TF, Chen JC, Ng CJ. Tuberculous peritonitis: Analysis of 211 cases in Taiwan. Dig Liver Dis 2012;44:111-7.
Chalya PL, Mchembe MD, Mshana SE, Rambau PF, Jaka H, Mabula JB. Clinicopathological profile and surgical treatment of abdominal tuberculosis: A single centre experience in Northwestern Tanzania. BMC Infect Dis 2013;13:270.
Niu B, Kim B, Limketkai BN, Sun J, Li Z, Woreta T, et al
. Mortality from spontaneous bacterial peritonitis among hospitalized patients in the USA. Dig Dis Sci 2018;63:1327-33.
Chow KM, Chow VC, Szeto CC. Indication for peritoneal biopsy in tuberculous peritonitis. Am J Surg 2003;185:567-73.
Thoreau N, Fain O, Babinet P, Lortholary O, Robineau M, Valeyre D, et al
. Peritoneal tuberculosis: 27 cases in the suburbs of Northeastern Paris. Int J Tuberc Lung Dis 2002;6:253-8.
Braun MM, Byers RH, Heyward WL, Ciesielski CA, Bloch AB, Berkelman RL, et al
. Acquired immunodeficiency syndrome and extrapulmonary tuberculosis in the United States. Arch Intern Med 1990;150:1913-6.
Rivara MB, Mehrotra R. The changing landscape of home dialysis in the United States. Curr Opin Nephrol Hypertens 2014;23:586-91.
Najafi I. Peritoneal dialysis in Iran and the Middle East. Perit Dial Int 2009;29 Suppl 2:S217-21.
Shen TC, Huang KY, Chao CH, Wang YC, Muo CH, Wei CC, et al
. The risk of chronic kidney disease in tuberculosis: A population-based cohort study. QJM 2015;108:397-403.
Lima NA, Stancic C, Vos D, Insua MM, Lima CC, de Castro RL, et al
. Hospital admissions for tuberculous pericarditis in the United States 2002-2014. Int J Mycobacteriol 2019;8:347-50.
] [Full text]
Glaziou P, Floyd K, Raviglione MC. Global epidemiology of tuberculosis. Semin Respir Crit Care Med 2018;39:271-85.
Migration Policy Institute. State immigration data profiles; 2017. Available from: migrationpolicy.org/programs/data-hub/state-immigration-data-profiles. [Last accessed on 2020 Feb 22].
Liao YJ, Wu CY, Lee SW, Lee CL, Yang SS, Chang CS, et al
. Adenosine deaminase activity in tuberculous peritonitis among patients with underlying liver cirrhosis. World J Gastroenterol 2012;18:5260-5.
DeShazo JP, Hoffman MA. A comparison of a multistate inpatient EHR database to the HCUP nationwide inpatient sample. BMC Health Serv Res 2015;15:384.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4]