A study on CABI score in correlation with comorbidity,microbiological profile and early mortality among pulmonary TB patients with bacterial co-infection

Abstract

Tuberculosis (TB)-related deaths are preventable, yet research on co-infections primarily focuses on human immunodeficiency virus (HIV)-positive patients, leaving a gap for HIV-negative individuals. Early clinical monitoring is vital for reducing mortality, but research lacks clear links between mortality scoring, microbiological data, and prognostic factors. Despite early mortality being absent among our participants, we found prognostic markers such as comorbidity, smear-negative TB, high bacterial load, and bacterial co-infection correlated with elevated CABI scores.

Keywords

Tuberculosis Asia diagnosis public health respiratory tract

Introduction

India stands as one of the countries with the highest burden of tuberculosis (TB), demonstrating an urgent need for concerted efforts to combat this public health challenge.¹ The Government of India has initiated the National Tuberculosis Elimination Programme (NTEP), intending to eradicate TB by 2025, surpassing the global timeline by 5 years. Despite newer modalities for the diagnosis and treatment of TB, it remains one of the top 10 fatal infectious illnesses, second to human immunodeficiency virus (HIV).² By 2025, the NTEP aims to achieve a consistent decline in the burden of TB, as well as related morbidity and mortality, while progressing towards the elimination of TB in India.^3,4

Every death attributed to TB represents a preventable tragedy. Infection with TB raises susceptibility to other infections increases, and lowers, life expectancy. In countries with a high incidence of TB, the majority of related fatalities occur within the initial eight weeks of treatment. These deaths predominantly affect individuals aged 15–49 years, particularly the poor, individuals living with HIV, and those with a low body mass index (BMI). Most deaths from TB are primarily due to advanced disease. These conditions are often further complicated by concurrent health concerns, including HIV, secondary infections (fungal, bacterial or bloodstream-related infections), and varying levels of undernutrition, ranging from moderate to severe. The majority of early mortalities among TB patients, regardless of whether they are also infected with HIV, can be directly linked to TB itself.⁵

Much literature suggests that patients diagnosed with TB require close monitoring and effective clinical care during the initial weeks of treatment, whether in clinical or programmatic settings, to minimise the risk of mortality. An autopsy-driven research revealed that almost half of the TB patients had bacterial co-infections.⁶ A study conducted in 2018 among HIV-negative PTB cases concluded that the bacterial co-infections increased the risk of early mortality up to 1.7 times.⁷ Patients with active TB may develop sepsis, as a consequence of concurrent bacterial infection or due to TB itself, which has been described as Mycobacterium tuberculosis-related septic shock.⁸ This syndrome is associated with the risk of an increase in levels of serum lactate, features of disseminated bacterial infection, and multi-organ failure.¹ These facts were evident from a study that reported that critically morbid hospitalised patients with PTB could potentially respond positively to broad-range antimicrobial treatment to reduce their mortality.²

A simple prognostic score (CABI) was developed in 2013. The clinical form of tuberculosis (C), age (A, years), adjusted body mass index (B, BMI, kg/m²), and status of HIV (human immunodeficiency virus) infection (I) were important predictors in the final model (p < 0.0001), which was the mortality risk.⁹ The optimal absolute risk threshold was determined to be 4.8%, yielding an 81% sensitivity and a 67% specificity. A recent study in 2022 used CABI scores in TB patients undergoing treatment. In their cohort study, the diagnostic performance of the CABI score was moderate. However, adding a new T for type of patient, A for age, B for BMI, and I for HIV infection (TABI) score, they obtained a good diagnostic performance with good specificity (60%) and good sensitivity (80%).¹⁰

Most studies on TB co-infections focus on HIV patients, leaving a gap in research for HIV-negative individuals. Moreover, there is a deficiency of correlation between mortality risk scoring, microbiological data and identifiable prognostic factors.

Material and methods

Ours is a cross-sectional observational study conducted between May 2023 and November 2024. A total of 200 newly diagnosed PTB patients (confirmed by CBNAAT) visiting the Directly Observed Therapy Short Course Chemotherapy (DOTS) Centre of the University College of Medical Sciences and Guru Teg Bahadur Hospital, Delhi, were recruited. The inclusion criteria were treatment-naive PTB and a patient's age >18 years. The exclusion criteria were patients critically ill, pregnant, or with HIV.

Ethical clearance for the study was granted by the IEC-HR at the University College of Medical Sciences (IECHR-2023-59-79), and informed consent was obtained from all participants. Patient details, including tuberculosis classification (C), age (A), adjusted BMI (B, BMI in kg/m²), and HIV status (I), were recorded for the CABI scoring system, which predicted mortality risk using Table 1.⁹ Two sputum samples were collected from each participant and subjected to bacteriological and mycobacterial processing, including bacteriological culture and identification, Ziehl-Neelsen (ZN) staining, and culture on Löwenstein-Jensen media for confirmation of MTB using the MPT 64 kit was performed on the isolates.¹¹ Study participants were followed up at 2 months to assess early mortality. Data were coded in MS Excel and analysed using SPSS 20.0, applying descriptive statistics and visualisation methods based on data type, with parametric and non-parametric tests for group comparisons, chi-squared or Fisher's exact tests for categorical data, and Pearson and Spearman's correlations for relationship assessments, with significance set at p < 0.05.

Table 1.

CABI scoring for a death prediction score in tuberculosis patients.

Age (years)		Adjusted BMI (kg/m²²)		To convert the point score into risk of death, look at the correspondence below
Range	points	Range	Points
< 20	0	<=10	50
[20–25]	2	[10–15]	44	Total points	Risk of death
[25–30]	3	[15–20]	38	<=20	<=0.1%
[30–35]	5	[20–25]	31	36	1%
[35–40]	6	[25–30]	25	40	2%
[40–45]	8	[30–35]	19	43	3%
[45–50]	10	[35–40]	12	45	4%
[50–55]	11	[40–45]	6	47	5%
[55–60]	13	>45	0	52	10%
[60–65]	14			55	15%
[65–70]	16	Clinical forms of TB		57	20%
[70–75]	17	Forms	Points	59	25%
[75–80]	19	PTB+	0	61	30%
[80–85]	21	ETB	4	62	35%
85+	22	PTB−	8	64	60%
				65	45%
		Sum up the points across all four factors above to obtain the total point score		67	50%
HIV status				68	55%
Status	Points			69	60%
Negative	0			71	65%
Positive	9			72+	>=70%

CABI: Clinical forms of tuberculosis, Age, Body mass index and Infection; HIV: human immunodeficiency virus; BMI: body mass index; PTB: pulmonary tuberculosis; ETB: extra tuberculosis.

Results

The study included participants aged 18–88 years, with a mean age of 33.7 years. There were 113 males (56.5%) and 87 females (43.5%). The 20–30 age group had the highest frequency of participants, followed by 18–20, 30–40, and 40–60 years, with fewer patients over 60 years, as depicted in Figure 1.

Of the 200 study participants, 163 had no comorbidity, while 37 had existing comorbidity, including diabetes, hypertension, COPD, cardiovascular diseases (CVDs) and stroke (Fig. 2). Since four patients had two comorbid conditions together, the total comorbidity amounted to 41 factors. Further, 16% presented with a contact history of TB.

The mean BMI of the total study participants was calculated to be 20.0 ± 4.4 kg/m², the median BMI was 20 kg/m², and the BMI ranged from 11 to 32 kg/m². Figure 3 shows the actual distribution.

The death risk score was calculated using data collected at diagnosis, which was then categorised into different levels of risk for death as described in Figure 4.

Out of 200 samples collected, 72% tested positive for ZN staining. The distribution of the study population based on various ZN grading revealed that the highest frequency of patients was smear-negative, followed by those with a grade of 2+, as shown in Figure 5.

Culture on LJ media resulted in growth in 144 samples, while 46 samples showed no growth, and 10 samples were contaminated during incubation. The mean time to culture positivity was 4.6 weeks. Significant bacterial growth in culture was seen in 19.5%. All participants in the study were monitored for mortality during the two months following the initiation of anti-TB treatment, but since no deaths were recorded, the early mortality rate was 0%.

Association of the levels of CABI scoring with demographic data

Non-parametric tests (Spearman correlation) showed a weak positive correlation between age and CABI scoring (rho = 0.07, p = 0.360). The Wilcoxon-Mann-Whitney U test revealed that the female participants had slightly higher average scores than males, as described in Figure 6.

The patients with comorbidity were associated with high levels of CABI scoring, which was statistically significant. The study participants with COPD and diabetes exhibited high levels of CABI scoring. A contact history of TB was found to be associated with CABI scoring. Spearman correlation indicated a moderate, statistically significant negative relationship between BMI (kg/m²) and CABI scoring (rho = −0.53, p < 0.001). For every 1 unit increase in BMI, CABI scoring decreases by 0.76 units (Fig. 7).

Association of various mycobacterial parameters with CABI scoring

The CABI scoring showed a significant correlation with the ZN smear grading and the time to culture positivity (Table 2). A significant negative correlation between CABI scores and time-to-culture positivity was seen, suggesting that as the time-to-culture positivity decreases, CABI scores increase.

Table 2.

Association of various mycobacterial parameters with CABI scoring (n = 200).

Sputum microbiology	CABI scoring	p-value
ZN smear grading***		<0.001^a
0	47.39 ± 5.15
1+	39.05 ± 5.56
2+	41.04 ± 7.19
3+	41.58 ± 6.23
Scanty	41.00 ± 3.25
Time to culture positivity (weeks)***	Correlation coefficient (rho) = −0.21	0.004^b

CABI: Clinical forms of tuberculosis, Age, Body mass index and Infection; ZN: Ziehl-Neelsen.

Kruskal-Wallis test.

Spearman correlation.

*** Significant at p < 0.05.

Association of bacterial co-infection with CABI scoring

Patients with bacterial co-infections had significantly higher median CABI scores (Wilcoxon-Mann-Whitney U test) as shown in Table 3.

Table 3.

Association of bacterial co-infection with CABI scoring (n = 200).

CABI scoring	Bacterial co-infection		Wilcoxon-Mann-Whitney U test
CABI scoring	Present (39)	Absent (161)	W	p-value
Mean (SD)	46.23 (9.01)	41.62 (5.63)	4541.500	<0.001
Median (IQR)	48 (43.5–51)	41 (38–46)
Min–max	13–62	28–67

CABI: Clinical forms of tuberculosis, Age, Body mass index and Infection; SD: standard deviation; IQR: interquartile range.

Association of CABI scoring and early mortality

Our study observed no deaths within 2 months of initiation of ATT in any of the study participants. Hence, the correlation of CABI scoring with early mortality was not possible.

Discussion

The age distribution of our study population aligns with national TB reports, which indicate that the highest number of cases is reported among individuals aged 25–34 years. A low notification rate was also observed among the elderly population aged >60 years.⁴ High notification rates among adults could be due to an increased incidence of the disease due to more socialisation, habits such as alcoholism and smoking, and their health-seeking behaviour. The elderly population adds up to a low notification rate also due to social immobilisation and dependency on others for seeking health care.^12–14 The male predominance in our study is similar to the shown predominance of male participants, which again aligns with recent national data, which reported that 61% of the notified TB cases were males.¹⁵ Such differences in the notification of cases may be attributed to various socio-economic and cultural reasons in low and middle-income countries, where females are economically dependent on males. It may also be due to the lower incidence of disease among females due to hormonal influence during the reproductive age group, or increased levels of CD4+ lymphocytes among females than males.^16,17

Comorbidity significantly influences the activation and progression of various communicable diseases, especially TB, predominantly diabetes and COPD, as found elsewhere.¹⁸ National reports suggest that diabetes is the most prevalent comorbidity among patients diagnosed with tuberculosis, next to HIV co-infection. Thus, NTEP collaborates with different public health programmes to design and implement interventions to address important comorbidity in TB patients.¹⁵ In fact, one in four newly diagnosed TB patients had a household member with a history of TB, and one in 10 had experienced a TB-related death in their household.¹⁹

Evidence suggests that for every one-unit reduction in BMI, the risk of acquiring TB increases by c. 14%, and the risk of relapse also increases by four times. In our study, the majority of the patients were underweight BMI <18 kg/m².⁴, not dissimilar to another study that included 9721 participants.²⁰

The majority of our patients were classified in the low-risk category with scores <5. Previously, a study reported the distribution of tuberculosis as 45%, 16%, 30%, and 9% in the categories of low, moderate, high, and critical risk of mortality by the same death risk scoring.⁹ This disparity may be attributed to several factors, including the exclusion of HIV-positive individuals from our analysis. Also the majority of participants in our study were adults, which significantly contributed to a reduction in the overall death risk score.

We found a link between the presence of bacterial co-infection and higher CABI scores, worsening outcomes, increasing treatment failure or mortality risk. We found no mortality, defined as death occurring within 8 weeks of the initiation of TB therapy. Others have found early mortality wiuthin 2-4 weeks varying from 3.8-28%, depending on the patients’ presentation at the time of diagnosis.^21–25 Factors reported to be contributing to mortality included severe illness at the time of presentation, extremes of age, and comorbidity such as HIV and TB retreatment status, none of which were present among our study participants.^26,27 Furthermore, owing to the excellent efforts of the NSP from 2017 to 2025, the NTEP has achieved a reduction of 18% in the overall mortality due to TB since 2015. This may also explain why our study did not observe any early mortality among TB patients.⁴

Although this limited our ability to evaluate the full utility of the CABI scoring system, we identified TB, smear-negative TB disease,⁹ a shorter time-to-culture positivity on the LJ medium, and the presence of bacterial co-infections as major factors

Overall, these findings suggest a multifactorial relationship between CABI scores and various clinical or microbiological indicators of TB severity. A high early mortality was found mostly in hospitalised patients with severe disease requiring respiratory support.^7,28,29

Footnotes

Acknowledgements

HOD Microbiology for departmental support and Dr Puneeta Hyanki for help at DOTS Centre.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

ORCID iDs

Bhavya R

Bineeta Kashyap

Rajat Jhamb

Ashwani Khanna

References

Kethireddy

Light

Mirzanejad

, et al. Cooperative antimicrobial therapy of septic shock (CATSS) database group. Mycobacterium tuberculosis septic shock. Chest 2013; 144: 474–482.

Schutz

Barr

Andrade

, et al. Clinical, microbiologic, and immunologic determinants of mortality in hospitalized patients with HIV-associated tuberculosis: a prospective cohort study. PLoS Med 2019; 16: e1002840.

Global tuberculosis report 2023 . Who. int. [cited 2024 Nov 6]. https://who.int/teams/global-tuberculosis-programme/tb-reports/global-tuberculosis-report-2023.

India TB report 2023 . Gov. in. [cited 2024 Dec 6]. https://tbcindia.mohfw.gov.in/2023/06/06/india-tb-report-2023/.

Bhargava

. Tuberculosis deaths are predictable and preventable: comprehensive assessment and clinical care is the key. J Clin Tuberc Other Mycobact Dis 2020; 19: 100155.

Karat

Omar

von Gottberg

, et al. Autopsy prevalence of tuberculosis and other potentially treatable infections among adults with advanced HIV enrolled in out-patient care in South Africa. PLoS One 2016; 11: e0166158.

Shimazaki

Taniguchi

Saludar

NRD

, et al. Bacterial co-infection and early mortality among pulmonary tuberculosis patients in Manila, the Philippines. Int J Tuberc Lung Dis 2018; 22: 65–72. PubMed.

May

Mya

, et al. Bacterial co-infection in pulmonary tuberculosis. Myanmar Med J 2019; 3: 9–14.

Pefura-Yone

Balkissou

Poka-Mayap

, et al. Development and validation of a prognostic score during tuberculosis treatment. BMC Infect Dis 2017; 17: 1–9.

10.

Kuotu

Diallo

, et al. External validation and improvement of the CABI prognosis score in Guinean patients treated for tuberculosis in Conakry. Pan African Med J - One Health 2022; 7: 6.

11.

Global tuberculosis report 2018 . Who.int. World Health Organization; [cited 2024 Oct 7]. https://www.who.int/publications/i/item/9789241565646 (2018).

12.

Dodd

Yuen

Sismanidis

, et al. The global burden of tuberculosis mortality in children: a mathematical modelling study. Lancet Glob Health 2017; 5: e898–e906. PubMed.

13.

Graham

Marais

Amanullah

. Tuberculosis in children and adolescents: progress and perseverance. Pathogens 2022; 11: 392. PubMed.

14.

Shariefuddin

Sari

Pandji

. Comparison of life quality between diabetic and diabetic with tuberculosis patients. Althea Med J 2016; 3: 440–445. PubMed.

15.

India TB report 2024 . Gov.in. [cited 2024 Dec 7]. https://tbcindia.mohfw.gov.in/2024/10/11/india-tb-report-2024/ World Health Organization (WHO). Global Tuberculosis Report 2021. Geneva, WHO, 2021. www.who.int/publications/i/item/9789240037021.

16.

Whitacre

Reingold

O'Looney

. A gender gap in autoimmunity. Science 1999; 283: 1277–1278.

17.

Mair

Hawes

Agne

, et al. Factors associated with CD4 lymphocyte counts in HIV-negative Senegalese individuals. Clin Exp Immunol 2008; 151: 432–440.

18.

Bhattacharya

Talukdar

Barman

, et al. Clinical spectrum and medical comorbidities in tuberculosis: a hospital-based study in Northeast India. Cureus 2020; 12: e10580.

19.

Shewade

Gupta

Satyanarayana

, et al. History of household member with tuberculosis or related death in newly diagnosed patients in India. Public Health Action 2020; 10: 53–56.

20.

Min

Kim

, et al. Effects of underweight and overweight on mortality in patients with pulmonary tuberculosis. Front Public Health 2023; 11: 1236099.

21.

Bhutia

Adhikari

. Pulmonary mycoses among the clinically suspected cases of pulmonary tuberculosis. Int J Res Med Sci 2015; 3: 260–268.

22.

Birlie

Tesfaw

Dejene

, et al. Time to death and associated factors among tuberculosis patients in Dangila Woreda, Northwest Ethiopia. PLoS One 2015; 10: e0144244.

23.

Oliveira

Silveira

JTPD

Beraldi-Magalhães

, et al. Early death by tuberculosis as the underlying cause in a state of Southern Brazil: profile, comorbidities and associated vulnerabilities. Int J Infect Dis. 2019;80: S50–S57.

24.

Field

Lim

Murray

, et al. Timing, rates, and causes of death in a large South African tuberculosis programme. BMC Infect Dis 2014; 14: 3858.

25.

Adamu

Gadanya

Abubakar

, et al. High mortality among tuberculosis patients on treatment in Nigeria: a retrospective cohort study. BMC Infect Dis 2017; 17: 170.

26.

Shewade

Nagaraja

Vanitha

, et al. Screening for severe illness at diagnosis has the potential to prevent early TB deaths: programmatic experience from Karnataka, India. Glob Health Sci Pract 2022; 10: e2100736.

27.

Jonnalagada

Harries

Zachariah

, et al. The timing of death in patients with tuberculosis who die during anti-tuberculosis treatment in Andhra Pradesh, South India. BMC Public Health 2011; 11: 921.

28.

Iliyasu

Mohammad

Yakasai

, et al. Gram-negative bacilli are a major cause of secondary pneumonia in patients with pulmonary tuberculosis: evidence from a cross-sectional study in a tertiary hospital in Nigeria. Trans R Soc Trop Med Hyg 2018; 112: 252–254.

29.

Attia

Pho

Nhem

, et al. Tuberculosis and other bacterial co-infection in Cambodia: a single center retrospective cross-sectional study. BMC Pulm Med 2019; 19: 60.