Pulmonary and meningeal tuberculosis patterns in children at King Abdul-Aziz university hospital in Jeddah, Saudi Arabia: A retrospective study

Background: Paediatric Tuberculosis (TB) is a significant health problem worldwide, with the World Health Organization (WHO) reporting almost 550,000 children infected with TB yearly. Several factors have affected the TB infection rate in Saudi Arabia, including Hajj and Umrah (Islamic pilgrimages to Mecca), travel, global migration, indigence, inaccessible healthcare services and drug resistance. 
Objectives: This retrospective study aimed to describe the patterns of pulmonary TB (PTB) and tuberculous meningitis (TBM) among children admitted to the Paediatric Department at King Abdul-Aziz University Hospital (KAUH) in Jeddah, Saudi Arabia, between January 2010 and May 2015. 
Methods: The patients’ clinical information was retrieved retrospectively from the hospital TB register and case notes, including the age, gender, clinical features, investigations, treatments and prognosis.


Introduction
Paediatric tuberculosis (TB) is one of the main health problems worldwide. The World Health Organization (WHO) has reported that almost nine million people are infected with TB each year, and nearly 550,000 (6.5%) of these cases are children [1,2]. In the developed countries, 3-6% of the total number of TB cases involve children, while in developing countries the prevalence becomes higher, reaching 25% paediatric cases [1,3]. Due to the diffi culties in isolating the Citation: Fida  mycobacterium responsible for TB infections in children, there has been a relative degree of negligence in the paediatric patient population [2,4,5].
Mycobacterium tuberculosis is the primary human pathogen that causes TB [1]. This disease damages the lungs, central nervous system, lymphatic system and circulatory system [1]. However, many primary pulmonary TB (PTB) patients are asymptomatic. Hilar adenopathy is a common radiological fi nding, while pulmonary infi ltrates in the mid and lower lung fi elds are rare [6,7]. Infrequently, a child presents with a high fever and cough, and symptoms similar to fl u symptoms, which improve within approximately one week [8,9]. Pulmonary manifestations are common among most children with TB, with only 25-35% of children developing extra pulmonary (EPTB) manifestations. The fi rst and most common presentation is lymphatic, which occurs in 66.7% of TB cases, while the second is meningeal, which occurs in 13% of the EPTB cases [10]. Tuberculous meningitis (TBM) is the most severe form, and carries a high mortality [11]. Approximately 1% of all TB cases develop central nervous system TB [12].
Although a TB infection can spread to any site, TBM is one of the most common results of PTB, developing 3-6 months after the primary infection [13]. The clinical signs of TBM begin with 2-3 weeks of prodromal malaise, headaches, a low-grade fever and personality changes. These are followed fi rst by a meningitis phase that mimics bacterial meningitis (fever, nuchal rigidity and an altered mental status), and then by a paralytic phase characterized by the rapid progression to stupor, coma, seizures, paralysis and death [13]. In 50% of the TBM cases there is a risk of serious long term neurological consequences or death [14]. The previous literature has reported mortality rates ranging from 30% to 90% [14].
One situation that increases the risk of EPTB is an advanced case of Human Immunodefi ciency Virus (HIV). The HIV pandemic, particularly in the advanced phases of this immune disease, has increased the overall burden of TBM [12], with high mortality rates reaching 67% (without HIV the mortality rate is 25%) [12,15].
The rates of TB vary across different regions of Saudi Arabia. For instance, compared to 32 cases per 100,000 in the capital, Riyadh, the infection rate has been found to be as high as 64 cases per 100,000 in Jeddah, which is a major hub of air and sea travel for pilgrimages to Mecca [16][17][18]. Two special factors may infl uence the infection rates in Saudi Arabia: the large number of emigrant workers who live in the country, and the many religious visitors who come to Mecca and Medina for Hajj and Umrah. Many of these labourers and visitors come from TB endemic countries, and previous research has found that more than 60% of the genotyped isolates come from imported strains [16][17][18]

Methods
This retrospective study described the patterns of PTB and TBM in children admitted to the emergency room (ER) and paediatric units at King Abdul-Aziz University Hospital, a large tertiary public hospital in Jeddah, Saudi Arabia, from January 2010 to May 2015. The patients included in the study were < 18 years old and treated for PTB and/or TBM. The patients' clinical information was retrieved from the hospital TB register and case notes. This information included the age, gender, hospitalization, clinical features, results of routine laboratory investigations, results of Cerebrospinal Fluid (CSF) analyses, chest radiographic fi ndings, lymph node biopsy results, treatments received, duration and prognosis.
The TB diagnosis protocol in our hospital is as follows: a standard Tuberculin Skin Test (TST) (5 TU in 0.1 ml; Radiant Parenterals, Ltd., Vaghodia, India) administered during the initial visit and read at 48-72 hours; a result of ≥ 5 mm induration was considered positive for TB [3]. Gastric aspirates (up to three consecutive samples), a lymph node biopsy, CSF analysis and culture, and pleural fl uid analysis and culture were obtained for the mycobacterial culture to confi rm the diagnosis. A diagnosis of PTB and/or TBM should be considered when a patient with the clinical symptoms or when the laboratory and radiological fi ndings lead to the suspicion of PTB with meningeal involvement. These clinical symptoms include fever, headaches, seizures and, upon examination, focal signs, neck stiffness and/or altered mentation. There should be evidence of space occupying lesions and/or signs of a raised Intracranial Pressure (ICP) before performing a Lumbar Puncture (LP) for a CSF analysis. TBM was suspected when the CSF showed high protein levels, low glucose levels, a high White Blood Cell (WBC) count, the absence of Cryptococcus spp. (India ink staining), and the absence of bacteria that commonly cause meningitis (gram staining) or their antigens (Streptococcus pneumonia, Neisseria meningitis, Haemophilus infl uenzae). TBM was confi rmed when M. tuberculosis was found in the CSF analysis. PTB was suspected in those cases with a history of fever and cough. The chest X-rays and chest Computed Tomography (CT) scans were reviewed for suspicious lesions (PTB), and three consecutive sputum samples were collected on different days for examination via Ziehl-Neelsen staining. Individuals were included in this study if suffi cient information was available to make a fi rm diagnosis, and were treated for PTB and/or TBM.

Statistical analysis
The collected data were analysed using the SPSS statistical software package, version 20. The parametric data were expressed as the mean and standard deviation (minimum and maximum), and the non-parametric data were expressed as numbers (percentages). The PTB and TBM comparisons were made using an unpaired Student's t test for the parametric parameters and a chi squared test for the non-parametric parameters. A cross tabulation between the PTB and TBM cases was conducted, and a P value < 0.05 was considered to be statistically signifi cant.
The mean CSF cell count was 35.0±24.0 (5.0-66.0). The mortality rate at the time of the study was lower in the PTB patients than in the TBM patients, with a signifi cant difference (15.0% versus 37.0%, P=0.05) ( Tables 2,3). With respect to the chest imaging in the PTB patients, a chest X-ray was performed  Data were expressed as mean [SD] (minimum -maximum) or number (%) as appropriate. Signifi cance was made between two groups using unpaired student "t" test. in 98.0% of the cases and a chest CT was performed in 39.0%. In the TBM patients, a chest X-ray was performed in 100% of the cases, a brain CT was performed in 96.0% and a brain MRI was performed in 86.0% (Table 3).
With regard to the drug administration, rifampin was given to 98.0% of the patients, isoniazid to 100.0%, pyrazinamide to 95.2%, ethambutol to 63.50%, streptomycin to 22.2% and dexamethasone to 23.8%. The duration of each antibiotic followed the international guidelines. In the treatment regimen, the number of patients given dexamethasone and streptomycin was signifi cantly lower in the PTB versus TBM patients (

Case 1 (TBM):
A 9-month-old unvaccinated Somali girl presented with fever and convulsions. Upon examination, she was hypertensive. The CSF lab results showed high protein (3.91 g/l), 100 cells (78% lymphocytes) and a negative culture. The staining and culture of the Acid-Fast Bacilli (AFB) were negative. The chest CT fi ndings showed upper lobar pneumonia associated with multiple mediastinal and right hilar lymphadenopathies. The brain CT showed acute communicating hydrocephalus. The brain MRI demonstrated diffuse leptomeningeal enhancement and thickening representing meningitis, with a distribution suggestive of TB.

Case 2 (PTB):
A seven-year-old Chadian girl presented with a history of fever over the previous 2 months, associated with night sweating and chills, weight loss, dry cough and a positive history of haemoptysis. The AFB sputum culture was positive. Her chest CT showed multiloculated patchy consolidations, one of which fi lled the entire upper right lobe showing multiple air bronchograms with multiple cavitations.

Discussion
In this study, the mean age of the patients was 6.28 years old, with a range of 0.1-14.0 years, which was consistent with previous studies indicating that age and immunity are the two major risk factors in the development of TB [7,9,19,20]. Fourteen out of 63 of these patients (22.0%) were exposed to TB over the previous two years. In this respect, it has been reported that TB is transferred from an infected person to a child via airborne particles that remain in the air for several hours [1,19,21].
In this retrospective study, the clinical presentation upon admission consisted most often of a fever (75.0%) followed by coughing (65.9%) in the PTB patients, and a fever (89.5%) followed by neurological disorders as convulsions (36.8%), disturbed consciousness (21.1%), neck stiffness (5.3%), and headache (5.3%) in the TBM patients. Previous studies have reported that a non-productive cough and mild dyspnoea are the most frequent symptoms in paediatric TB patients. In addition, systemic complaints, such as anorexia, fever and night sweats, were less commonly reported. Respiratory symptoms were found to be more common in infants and young children [1,22].
A major challenge of paediatric TB is providing an accurate diagnosis. For this research, TB was most often diagnosed via a positive TB culture, followed by a positive PPD. Additionally, it has been reported that less than 15.0% of children with TB had positive sputum smears, with mycobacterial cultures detecting the bacilli in 30-40% of the cases [1,23]. All of the guidelines agree that the isolation of the M. tuberculosis complex from different specimens is the gold standard for the diagnosis of TB [3]. However, in those cases with no bacteriological confi rmation, a childhood diagnosis in TB endemic countries should be based on a positive family history of TB, a positive tuberculin skin test, the presence of physical signs and symptoms and/or abnormal fi ndings on the chest X-ray [7,23,24].
There is a general consensus that chest radiography is essential for the diagnosis of TB, and should be done for any child with suspicious symptoms [3]. Meanwhile, disagreement in chest radiographies fi ndings for TB screening in children had been reported previously [25,26] In this study, active PTB was diagnosed on the basis of a combination of epidemiological (e.g., exposure, travel to or residence in a high prevalence area, previous TB), clinical (e.g., cough lasting longer than 2-3 weeks, fever, night sweats and weight loss), radiographic (e.g., infi ltrates, fi brosis and cavitation), microbiological (e.g., positive sputum smear or culture) and histopathological (e.g., caseation granuloma) features. Those patients in whom the clinical suspicion of TB is strong on the basis of clinical criteria should undergo chest radiography [6,27,28].
The diagnosis of TBM demands a high level of suspicion, and the CSF analysis should show mononuclear pleocytosis, increasing protein levels (100-150 mg/dl), an increasing WBC count (100-150 cells/μl) and decreasing glucose levels (<45 mg/ dl) [7]. An exact diagnosis can be made via direct Ziehl-Neelsen staining (with low sensitivity) or CSF culture (too slow), which could delay the treatment decisions [12,29]. In the current study, the mean values were 1.36±1.08 (mg/dl) for the protein level, 2.73±1.58 (mmol/L) for the glucose level and 48.5±23.41 (cells/ul) for the cell count, which were consistent with those of other studies. An acid fast bacillus culture is routine practice, and should be done in suspicious cases of TB. However, due to logistic reasons and the long time period before the results become available, the majority of the diagnoses are based on either clinical suspicion associated with improvement due to empirical treatment, and indirect clinical, laboratory (specially CSF analysis) and radiological fi ndings that provide a level of suspicion of TBM [12,30]. In the study by Luma, et al. [12], headaches and fevers were the most common symptoms, involving 74.1% and 59.3% of the patients, respectively. We, therefore, suspected TBM in the presence of central nervous system (CNS) disease symptoms/signs, and the diagnosis relied almost entirely on the CSF fi ndings of high proteins levels, low glucose levels and leucocyte pleocytosis, especially where there was mononuclear cell predominance [13].
Citation: Fida  In this study, the TBM mortality rate was 26.3% because the delays in the diagnoses greatly increased the number of patients that died. Similar results were found in a Johannesburg study in which the authors reported an increasing in mortality rate (from 56.0% to 90.5%) when the treatment was delayed for longer than 24 hours [12,13,31]. Obviously, there is a need to fi nd ways to help provide an early diagnosis, especially in indigent settings [12]. The key strategy for reducing and preventing TB is the immediate and rapid recognition and remediation of TB patients, as well as patient education and vaccination [31].
TB treatment consists of two steps: an intensive step using a combination of antibiotics to kill the quickly growing bacilli, and a continuation step using less drugs to excise the slower growing incessant bacilli [19]. For TBM, there are two standard treatment plans: 1) two months of an isoniazid, rifampin, pyrazinamide and ethambutol combination, then four months of isoniazid and rifampin, and 2) 12 months of therapy, which has been recommended by several authors [12,19,32]. The use of steroids in TBM as an adjunctive treatment has been shown to effectively decrease the rates of death and severe disability [19,33,34]. With regard to the challenges of noncompliance for a number of patients, which affects the outcomes and causes treatment failure, a new strategy has been recommended: Directly Observed Therapy (DOT), particularly in the paediatric setting [3,28].
The current study showed that physicians do follow the guidelines. Previously, a retrospective study was carried out at the King Khalid National Guard Hospital in Jeddah, between June 1993 and June 1999, involving 147 patients with positive TB cultures. Only 126 completed treatment, and the others were considered to be treatment failures. The overall rate of success was 69.4%, while the rate of was failure was 30.6%, and the authors reported that the two factors which caused the failures were drug resistance and the lack of commitment [35]. Even though the National Tuberculosis Program (NTP) and DOT were implemented, a drug resistance situation decreased the WHO's target of healing 85% of TB patients. This was likely due to the lack of commitment from the patients and the availability of over the counter anti-TB drugs [16].

Study limitations
This study did have a number of limitations. For example, it was a retrospective study, which introduced a problem in the availability of a suffi cient level of information (missing data) about the patients, leading to the exclusion of otherwise qualifi ed cases. In addition, the small sample size restricted the statistical analyses. Moreover, most of the diagnoses were dependent on the clinical presentations and not the investigations, so many cases may have been missed. Additionally, treatment resistance could not be detected due to the retrospective nature of this study. Finally, only patients from one health centre were included, which makes the generalization of these results throughout the healthcare centres in Jeddah diffi cult. However, this study did describe and characterize the challenges faced by many clinicians in the diagnosis and management of PTB and TBM.

Conclusions
Two important factors that play large roles in the prevention and control of TB are the national TB program and the laboratories status. The results of the present study suggested that the diagnosis and treatment of children with TB are quite challenging, and that the tools currently employed in our facility are inadequate, leading to a high mortality rate, specifi cally among TBM cases. Effective management demands immediate recognition and treatment, especially in the paediatric population. The results of this study suggest that in order to increase the rate of early TB diagnoses, the diagnostic laboratory standards must be uniform across the country.