Outcome of active pulmonary tuberculosis patients requiring respiratory intensive care admission

Among hospitalized patients with active TB, the most common symptoms include fever, night sweats, weight loss, and cough [2]. In only one study, dyspnea was the most commonly reported symptom [3]. The mean symptom duration before hospital admission was ∼30 days in most of the studies. The presence of extrapulmonary TB ranged from 19 to 64% of the cases. Comorbidities, especially those related to immunosuppression, such as HIV infection, are considered risk factors for developing respiratory failure and requiring mechanical ventilation (MV). The most common radiological findings are reticular infiltrates and consolidation, and cavitations can occur in 27–50% of cases [2].


Introduction
Across the world, tuberculosis (TB) remains an important public health problem, especially in developing countries. One-third of the world's population is infected with Mycobacterium tuberculosis. Every year almost two million people die from TB, most of them in low-income and middle-income countries [1].
Among hospitalized patients with active TB, the most common symptoms include fever, night sweats, weight loss, and cough [2]. In only one study, dyspnea was the most commonly reported symptom [3]. The mean symptom duration before hospital admission was ∼30 days in most of the studies. The presence of extrapulmonary TB ranged from 19 to 64% of the cases. Comorbidities, especially those related to immunosuppression, such as HIV infection, are considered risk factors for developing respiratory failure and requiring mechanical ventilation (MV). The most common radiological findings are reticular infiltrates and consolidation, and cavitations can occur in 27-50% of cases [2].
The leading cause of respiratory intensive care unit (RICU) admission was respiratory failure, and Acute Physiology and Chronic Health Evaluation II (APACHE II) scores ranged from 13 to 23 in most of the studies [4]. Frame et al. [5] evaluated the factors associated with the development of respiratory failure and the need for MV. Gram-negative pneumonia or sepsis, chronic obstructive pulmonary disease, history of poor compliance with TB treatment, and cancer were predictors of respiratory failure. However, despite their high mortality rate, the TB-related critical conditions are rarely reported. The cases of TB requiring intensive care represent 1-3% of all patients with TB [5][6][7].
Limited studies reported a few factors that contribute to mortality among critically ill patients with TB, such as extensive fibrocavitary disease and consolidations on chest radiographs, acute respiratory distress syndrome, sepsis, and multiple organ failure (MOF) also carry a very high mortality [3,8,9]. Appropriate anti-TB treatment is an important factor that can affect patient outcome. Higher mortality is found among patients who do not receive optimal treatment [10].
There are limited data regarding patients with active pulmonary tuberculosis (APTB) requiring ICU admission, especially when reported from developing countries. To our knowledge, this has never been

Therapeutic characteristics
This included anti-TB drug regimen, oxygen therapy, and MV.

Case definitions of active pulmonary tuberculosis
(1) Smear and/or culture-positive pulmonary TB patient in one of three ways: patient with at least two sputum specimens positive for acid fast bacilli (AFB) by microscopy or patient with at least one sputum, bronchial aspirates, or bronchoalveolar lavage fluid positive for AFB by microscopy and radiographic abnormalities consistent with pulmonary TB and decision by a physician to treat with a full course of anti-TB chemotherapy or patient with at least one sputum specimen positive for AFB by microscopy, which is culture positive for M. tuberculosis [11]. (2) Histopathological diagnosis of computed tomography-guided biopsy or transbronchial samples showing caseating granulomatous reaction.

Drug-resistant tuberculosis
Drug-resistant TB is defined as multidrug resistant tuberculosis (MDR-TB) in which colonies are resistant to at least isoniazid and rifampicin [11].

Extensive drug-resistant tuberculosis
Extensive drug-resistant tuberculosis (XDR-TB) is defined as TB in which in addition to being resistant to rifampicin and isoniazid, colonies are resistant to one of the fluoroquinolones and one or more than one of the injectable drugs (kanamycin, amikacin, and capreomycin) [11].

Respiratory failure
Respiratory failure was defined as a PaO 2 , measured at rest at sea level, of less than 60 mmHg or a PaCO 2 above 49 mmHg. Type I respiratory failure was defined as hypoxemia without hypercapnia, but type II respiratory failure is associated with hypercapnia [13,14].

Multiple organ failure
MOF was defined as the failure of more than one organ.

Acute respiratory distress syndrome
Acute respiratory distress syndrome was diagnosed based on the criteria defined in the American-European consensus conference [15].

Early/late intensive care unit admission
Early ICU admission was defined as admitted directly or transferred to an ICU within 4 days of admission. studied before in Egypt. This study aims to describe and determine the clinical presentation, mortality rate, and risk factors associated with mortality in patients with APTB requiring RICU admission.

Patients and methods
This was a retrospective-prospective study design initiated in January 2011 in which adult patients (>16 years) with APTB previously or subsequently admitted to the RICU of Abbassia Chest Hospital between January 2009 and December 2010 (retrospective part) and between January 2011 and December 2011 (prospective part), respectively, were included.
Patients who stay or stayed at RICU for less than 24 h or presenting with inactive pulmonary TB or extrapulmonary TB were excluded.
The Abbassia Chest Hospital is a special, tertiary care hospital with 600-bed occupancy in which 100 of them are for TB, 32 ICU beds; three of them are isolated for APTB. TB treatment in the hospital is according to the Egyptian national guidelines [11].
The following data were obtained from the medical records:

Demographic characteristics
It included age, sex, special habits of medical importance, and comorbidities.

Clinical characteristics
(1) Pulmonary TB presentations: symptoms included the presence or absence of the following presenting symptoms: fever, cough, dyspnea, hemoptysis, loss of body weight, and symptoms more than 30 days. Signs included blood pressure, temperature, and heart rate. (2) Pattern of pulmonary TB: whether it is miliary or pneumonic TB type.

Laboratory investigations on admission
This included white cell count, hemoglobin, coagulation profile, liver and renal function tests, electrolytes, and arterial blood gases analysis.
The level of significance was considered at P value less than 0.05 as significant, otherwise nonsignificant.

Results
A total of 100 patients were included during the study period from January 2009 to December 2011 (retrospective part included 60 patients and prospective part included 40 patients). Of 100 patients studied, 74 (74%) died in the ICU.

Demographic characteristics
The baseline demographic characteristics of both survivor and nonsurvivor patients are shown in Table 1. The median age of all patients was 38 years. Men outnumbered women. Female frequency was significantly higher in nonsurvivors than in survivors. Liver impairment (48%) was the most common comorbidity. Four of the 100 patients were positive for HIV infection.

Clinical characteristics
The clinical characteristics including causes of RICU admission and complications of studied cases are illustrated in Tables 2 and 3. Fever was the most common symptom. Pneumonic pattern of pulmonary TB represented the majority of cases than miliary pattern. The majority of radiographic findings were moderate to far advanced lesions. The median length of stay in RICU for all patients was 5 days (range 2-60 days). Sixty-five percent of studied patients needed MV. Respiratory failure was the most common cause of admission in the RICU. Complications occurred in 45% of cases in which electrolyte disturbance was the most common complication recorded. Lower diastolic Late ICU admission was defined as transferred to an ICU after more than 4 days of admission to TB ward.

Glasgow Coma Scale
GCS score was calculated as previously stated in details using medical calculator available on the internet [16].

Acute Physiology and Chronic Health Evaluation II
APACHE II score was calculated using medical calculator available on the internet [17].

Statistical analysis
The collected data were coded, tabulated, and statistically analyzed using statistical package for social sciences software (SPSS, version 18.0; SPSS Inc., Chicago, Illinois, USA).
Descriptive statistics were presented for numerical parametric data as mean ± SD, minimum and maximum of the range, and for numerical nonparametric data as median and first and third interquartile range, whereas they were presented for categorical data as number and percentage.
Analysis was performed for quantitative variables using independent t-test in cases of two independent groups with parametric data and using the Mann-Whitney U-test in cases of two independent groups with nonparametric data. Inferential analysis was performed for qualitative data using the c 2 -test for independent variables.
Receiver-operator characteristic curve was used to evaluate the performance of different tests and to differentiate between certain groups. Logistic regression model was used to find out independent factors affecting nonsurvival. than in survivors. Total leukocytic count, international normalized ratio, urea, creatinine, and bilirubin were significantly higher in nonsurvivors than in survivors.
blood pressure, far advanced lesion, type II respiratory failure, higher APACHE II score, lower GCS score, more need for MV, and electrolytes disturbances complication occurrence were significantly more frequent in nonsurvivors than in survivors in the RICU. Respiratory alkalosis was the cause of RICU admission in only two survivor cases, and hepatotoxicity was a recorded complication in also two survivor cases only. Thus, both were significantly more in survivors than in nonsurvivors.

Laboratory investigations on admission
The laboratory investigations on admission of studied cases are shown in Table 4. Albumin, pH, and O 2 saturation were significantly lower in nonsurvivors   other factors such as female sex, far advanced lesion, electrolytes disturbance, and renal impairment on admission were nonpredictors of mortality with diagnostic accuracy less than 85% (Table 7). Albumin, O 2 saturation, GCS, urea, creatinine, and APACHE II score were nonpredictors for mortality as indicated by low area under curve (<0.9) ( Table 8).

Discussion
Across the world, TB remains an important public health problem, especially in developing countries. Despite the availability of curative therapy, a large proportion of patients with TB are being hospitalized.
In-hospital mortality rates remain high, particularly among patients with TB requiring ICU [8].
Pulmonary TB requiring ICU care is rare [5]. To our knowledge, this was the first Egyptian study to deal with APTB inside RICU. This study aimed to evaluate, on retrospective-prospective basis, the clinical presentation, mortality rate, and risk factors associated with mortality among RICU patients with APTB.  [5,19,21]. Sex distribution was in line with data from previous studies, as men (80%) were seen more often than women (20%) [10,19]. One possible explanation for this disparity might be the higher exposure of men to droplet infections due to a greater prevalence of outdoor occupations. Aggravating factors such as smoking, exposure to air pollutants, and industrial exposure may also predispose men to TB [19]. However, women were significantly higher in nonsurvivor than in survivor patients, and it was one of the risk factor of mortality. In contrast, Erbes et al. [10] found no significant

Therapeutic characteristics
Comparison between survivors and nonsurvivors with respect to treatment regimens was significantly different (Table 5).

Risk factors for nonsurvival
On studying different involved factors, pneumonia pattern, female sex, far advanced lesion (by radiology), and renal impairment were found to be a significant model for risk factors for nonsurvival (Table 6) using logistic regression model.

Predictive factors for respiratory intensive care unit mortality
On studying different involved factors to search for predictors of mortality, MV was the only predictor of mortality with diagnostic accuracy of 85%, whereas difference between survivors and nonsurvivors with respect to patient sex. This discrepancy may be related to difference in disease severity, comorbidities, and need for MV between both studies.
In the present study, a history of renal impairment was found to be a risk factor for ICU mortality, similar to the  SGOT, serum glutamate oxaloacetate transaminase; SGPT, serum glutamate pyruvate transaminase; TLC, total leukocytic count; a Mann-Whitney U-test; b Independent t-test; *Significant.    and hypocalcemia, especially in MDTR-TB. Electrolyte disturbances may be manifested as subtle complaints up to lethal cardiac arrhythmias [24]. This may explain the cardiac complications such as atrial fibrillation and ventricular fibrillation that occurred in the current study.

Variables Patients Nonsurvivors Survivors
Hypoalbuminemia was a common laboratory finding in previous studies dealing with TB patients requiring ICU admission similar to current study results [3,10,18]. Unlike previous studies, hypoalbuminemia had affected mortality among current study patients. However, it was not among the predictor of mortality.
Pulmonary TB is a rare primary cause of acute respiratory failure (ARF) [6]. However, high mortality rates have been reported in patients with ARF arising from TB [3]. In the present study, respiratory failure was clinically manifested in 74% of study population with a morality rate 75%, which lays in range with previous studies results [3,5,8,10].
Appropriate anti-TB treatment is an important factor affecting patient's outcome. It was found that a higher mortality is present in patients who did not receive an optimal treatment including isoniazid and rifampicin [10,23]. In the current study as well as in other studies, optimal treatment is difficult to achieve in those critically ill TB patients due to use of alternative regimens (without isoniazid and rifampicin) for maintenance therapy in patients with impaired liver function or organ dysfunction and with low serum concentrations of anti-TB drugs as results of uncertain enteral absorption of drugs and hypoalbuminemia [23,[25][26][27]. The combined MDR-TB and XDR-TB rate in the present study reached 4% with its subsequent expected adverse outcomes such as treatment failure and death.
MV was the only predictive factor associated with ICU mortality in the current study, which is in accordance with Lee et al. [3] and Erbes et al.'s [10] studies. The mortality rate in patients with APTB requiring MV in the current study was high (83%). ARF caused by TB and requiring MV has been associated with morality rates ranging from 17.5 to 81% [3,5,10,23] and being 93% in one study [8].
In the ICU setting, a high clinical suspicion is mandatory in suspected cases, with smear microscopy for AFB and culture of respiratory secretions in suspected cases. Biosafety measures to suspected or diagnosed cases including individual isolated negativepressure rooms, usage of endotracheal suctioning without disconnection (closed system), and a bacterial filter should be placed in the expiratory line of the study by Silva et al. [18]. However, it did not influence the mortality rates significantly. Erbes et al. [10] identified a history of chronic pancreatitis as independent risk factor for mortality, which was related to chronic alcohol abuse. Other comorbidities such as diabetes or even coinfection with HIV encountered in the current study as well as in other studies [3,10,18] did not influence the mortality rates significantly in patients with TB requiring ICU. In fact, mortality in patients with TB-HIV coinfection seems to be related to the patient's overall degree of immunosuppression [22].
The clinical presentation of current study population was similar to previous studies [3,10,19]. Fever was the most common symptom observed upon presentation to the hospital and was seen in all (100%) patients. Fever was indicative of a high-grade of infection that required immediate attention [19]. Dyspnea, cough, hemoptysis, and loss of body weight were also frequent presenting symptoms, similar to other previous studies [3,10,19].
In the current study, nonsurvivor patients with far advanced lesion were significantly more than survivor patients, and far advanced lesion was significant risk factor of mortality. Other studies reported increased mortality among those with extensive bilateral infiltrates and cavitary lesions, but it did not reach a significant value [10]. This may be due to the increased number of studied patients with extensive lesions in current study in comparison with Erbes et al.'s [10] study.
APACHE II score was found to be significantly higher in nonsurvivor patients compared with survivor patients. It has been previously found that APACHE II score at ICU admission might underestimate the morality rate among TB patients requiring MV [23]. In the present study as well as in Silva et al.'s [23] study, the mean APACHE II score on admission was 20.5 ± 5.9 and 21.2 ± 6.5, respectively, indicating a mortality rate around 30%. However, the mortality rates were found to be 74 and 67.8%, respectively [19]. It is possible that the events occurring during hospitalization, such as respiratory failure or any other complications, contributed to the increased mortality rates [23].
There was a significant increase in electrolyte disturbances as a recorded ICU complication in nonsurvivors compared with survivor patients. Electrolytes disturbances in TB patients are likely multifactorial including disease chronicity, poor nutritional status, anti-TB treatment-induced diarrhea and vomiting, and aminoglycosides and capreomycin-induced hypokalemia, hypomagnesemia, ventilator circuit. Healthcare professionals should wear N95 masks while performing the intubation/ suctioning procedure and should also be periodically screened for TB [28].
The current study as any other study has its own limitations that should be noted. Part of the information was obtained retrospectively from patient records and probably was not as complete and accurate as when data collection was performed in the prospective part of study. Singlecenter study with relatively small sample size may have lacked sufficient statistical power to reveal an association with some of the factors. Despite these limitations, the present results provide important implications for similar demographic areas and clinical settings.
In conclusion, the present study found a very high mortality rate among APTB patients requiring RICU admission. Risk factors identified for nonsurvival were pneumonia pattern and far advanced lesion by radiology, female sex, and renal impairment. Only MV was a predictor of RICU mortality. Measures aimed at promoting appropriate management of these recognized risk factors and predictor of morality could contribute to improving outcome.