Population-Based Tuberculosis Disease Prevalence Survey in Ghana: The Role and Lessons Learnt from the Laboratory

Background: Bacteriologically-confirmed tuberculosis (TB) cases used in calculating TB prevalence in a country are obtained through laboratory examination of sputum specimens. Objective: This article describes laboratory processing of specimens, results overview, conclusions and key lessons learnt from the perspective of laboratory personnel involved in the conduct of TB disease prevalence survey in Ghana in 2013. Methods: Symptoms screening and Chest X-ray suggestive of TB were used to select participants who produced sputum to confirm TB cases using microscopy, culture and Xpert MTB/RIF assay (GeneXpert). Results: A total of 15,935 single and paired sputum specimens were received from eligible participants. About half of Ziehl-Nielsen (129/263) and Auramine O (122/246) stained smear positives were scanty positive. Culture positivity rate for Mycobacterium tuberculosis complex was 266/14,994 (1.7%) and 100/15,179 (0.7%) in Mycobacterial Growth Indicator Tube (MGIT) and Lowenstein-Jensen (LJ) media respectively; while non-tuberculous mycobacterium was 294/14,994 (1.96%) and 167/15,179 (1.1%). Total contamination rates in MGIT (5.4%) were higher than in LJ (1.7%). Prevalence of smear positive TB and bacteriologically confirmed TB among adult population (≥15 years) was estimated at 111 (95% CI: 76 145) and 356 (95% CI: 288 425) per 100,000 population respectively. Conclusions and Lessons Learnt: Direct supervision of specimen collection by well-trained laboratory personnel, timely transportation of specimens from field to laboratory, prompt specimen processing and use of electronic data management systems are essential for a reliable TB disease prevalence How to cite this paper: Addo, K.K., Addo, S.O., Bonsu, C., Mensah, E., Edusei, S., Dedzo, P., Omari, M.A., Kudzawu, S., Ganu, H., Atiadevie, S.K. and Bonsu, F.A. (2019) Population-Based Tuberculosis Disease Prevalence Survey in Ghana: The Role and Lessons Learnt from the Laboratory. Journal of Tuberculosis Research, 7, 95-108. https://doi.org/10.4236/jtr.2019.72009 Received: May 22, 2019 Accepted: June 27, 2019 Published: June 30, 2019 Copyright © 2019 by author(s) and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY 4.0). http://creativecommons.org/licenses/by/4.0/ Open Access


Introduction
Despite being declared a global health emergency in 1993 by the World Health Organization (WHO), tuberculosis (TB) remains a disease of public health importance globally infecting about one third of the world's population with about two million deaths occurring annually [1]. In Ghana, an estimated 14,668 (Pulmonary, bacteriologically confirmed-7682; Pulmonary, clinically diagnosed-5364; Extra-pulmonary-1181; Relapsed-441) cases were reported in 2014 [2].
While trend analysis of programmatic data provides useful information to indirectly assess progress, there are limitations. Firstly, we are unable to entirely rely on the routine surveillance system owing to its coverage, completeness and accuracy. Secondly, the current WHO estimates may be unrealistic; since the last well conducted prevalence survey based on TB disease in Ghana was in 1957. In the year 2000, TB burden measurement was done, but was based on TB infection, which is less reliable and not very accurate. As of 2012, TB case detection rate in Ghana was very low at 33% with estimated prevalence of more than 70 per 100,000 population [3]. With these aforementioned reasons, it was imperative for a prevalence survey to be conducted. Therefore, in 2013, a population-based TB disease prevalence survey was conducted with a general objective to obtain a direct measurement of the absolute burden of disease caused by TB in Ghana. A combination of symptom screening using standardized questionnaire and chest X-ray was used in order to selectively target the survey participants who provided sputum for bacteriological tests. We report here the general overview of the prevalence survey process with emphasis on the laboratory activities including sample collection and processing, data obtained, analysis and interpretation, as well as lessons learnt.

Study Design
The study was cross sectional and population based carried out on a random sample of the general adult population (15 years and above) in Ghana, in which the number of people with TB disease was measured. There were 98 survey clusters in two strata (urban-53, rural-45). k 2 (0.5); where k is the coefficient of between-cluster variation Response rate (85%).

Selection of Eligible Participants
All inhabitants including visitors in a cluster were enumerated by households prior to the survey. After enumeration, the survey team selected eligible participants from each household based on study inclusion criteria and provided invitation cards to participate in the survey.

Inclusion Criteria
All inhabitants of selected clusters who were 15 years old and above; residents and visitors who have lived in the household for the most of the last two (2) weeks preceding survey census (which means at least 7 days or more in the last 14 days); permanent residents who have lived in the household few days (less than 7 days) in the last 2 weeks before the census (i.e. residents who were away for some time, but that just came back home a couple of days before the census).

Exclusion Criteria
Eligible residents who are incapable of consenting (mentally incapacitated, people between 15 -17 years who cannot secure consent etc.); visiting residents who have lived in the enumeration area for less than 7 days in the last 2 weeks prior to census day; permanent resident who have not lived in the household in the last 2 weeks before the census (i.e. people who travelled since more than 2 weeks before the census); hotel guests; prisoners; students in hostels; residents of student dormitories in boarding school premises; residents of seminaries, sisters convents and monastries; diplomatic compounds; military facilities; street dwellers (with no fixed address).

Screening Strategy
Symptom screening and/or chest X-ray and/or chest X-ray exemption were used to identify individuals eligible to produce sputum for bacteriological examinations as recommended by WHO [4]. The overall algorithm for screening eligible participants and specimen processing is shown in Figure 1

Laboratory Technical Staff and Sample Processing Sites
Two groups of well-trained laboratory technicians and biomedical scientists were fully engaged for the study. The first group consisted of one technician attached to each of the four field survey teams to supervise the collection of sputum

Sputum Sample Collection and Processing
Two sputum specimen (spot and morning) were collected from each eligible participant according to symptoms screening or abnormal chest radiograph, and from all pregnant women who were exempted from chest X-ray. All the sputum specimens collected from participants in a particular survey cluster were trans-

Sputum Digestion and Decontamination
Sputum specimen were first processed using Mycoprep™ (BD Diagnostic System, Sparks, MD, USA) which is a commercially available Sodium hydroxide-N-acetyl L-cysteine (NaOH-NALC) formulation for sputum digestion and decontamination. This initial process was done according to manufacturer's instructions. In brief, equal volumes of specimen (3 -5 ml) and NaOH-NALC solution were mixed together and the preparation allowed to stand for 15 minutes at room temperature (25˚C). Next, about 40 -45 ml of phosphate buffer saline (PBS) solution with pH 6.8 was added to the decontaminated specimen and allowed to stand for 20 minutes to neutralize the reaction. Further, the whole preparation was centrifuged at 3000×g for 15 minutes to concentrate the specimen and also wash the NaOH solution. Then the supernatant was discarded to obtain sediment. Finally, a small volume (2 ml) of PBS (pH 6.8) was added to the sediment to make the inoculum for smears and cultures.

Smear Preparation and Microscopy
Two smears were prepared from each inoculum and stained for microscopy using ZN and Auramine O staining methods. The smears were examined after staining for the presence or absence of acid fast bacilli (AFB). The ZN stained smear was examined using Olympus™ light microscope (Olympus Corporation, Shinjuku, Tokyo, Japan) with oil immersion at 1000× and the Auramine O stained smear examined with Primo Star iLED™ microscope (Carl Zeiss Microimaging, Oberkochen, Germany), without oil immersion at 400×. Smears were graded using the World Health Organization and the International Union against Tuberculosis and Lung Diseases (WHO/IUALTD) standards. All smear positive specimens and contaminated culture (where sediments were available) were run on GeneXpert for Mycobacterium tuberculosis confirmation as well as rifampicin resistance. and another in the test zone (T).

Laboratory Case Definitions
Culture-confirmed TB case: isolation of MTBC from a sputum specimen grown on either LJ, MGIT or both.
Culture confirmed non-TB case: isolation of NTM from a sputum specimen grown on either LJ, MGIT or both. Laboratory and data analysis experts from WHO visited KBTH and NMIMR TB laboratories on three occasions for monitoring and evaluation.

Data Management and Analysis
All data collected from the field and laboratory were transferred electronically to a central server hosted at the secretariat of the National Tuberculosis Programme (NTP). Data analysis was performed using Stata (v12, Stata Corporation, USA) and SPSS (v20, IBM Corporation, USA).

Ethical Consideration
The study obtained ethical approval from the Institutional

Results
Among

Discussion and Lessons Learnt
The overall goal of the prevalence survey was to gain better understanding of the burden of TB disease and identify ways by which TB control can be improved.
The laboratory played a very important role in the successful conduct and outcome of the survey. Sample collection and processing were all performed by the laboratory which generated data for calculating the prevalence of TB disease.
The proportion of survey participants eligible for sputum collection was higher (13.4%) than what was reported in other African countries such as Ethiopia (13%) [5] and Nigeria (10.6%) [6], but lower than in Zambia (14.6%) [7] and Gambia (13.8%) [8]. The high sputum collection rate (98% for at least one specimen and 93% for both spot and morning specimen) was comparable to results from surveys in other countries [5] [7] [8]. In the case of Ghana, we can attribute this finding to factors such as effective community sensitization prior to commencement of the survey, professional skill of the survey team, supervision by field laboratory staff whereby in some instances, follow-up to residences of participants for specimen collection and the provision of digital chest X-ray examination. In addition, those who returned the early morning sputum sample were also given a vitamin fortified cereal.
Smear results from majority of the specimens were negative which was much expected in a TB prevalence survey where in addition to the gold standard muco-purulent sputum, salivary specimens are accepted [9]. It was also observed Journal of Tuberculosis Research that smear positivity were of the lower grade (scanty) and this could be due to the paucibacillary load of specimens received. Thus, the quality and bacillary load of sputum specimens collected as well as method used can determine AFB detection rate [10]. The high number of smear negative-culture positive speci- were observed during the initial stages of the survey period due to some operational challenges. The most common types of contamination were fungal and non-mycobacterial growth leading to liquefying or discolouration of inoculated LJ media and homogenous turbidity of MGIT culture. As expected, contamination rate of the MGIT were higher than LJ cultures. Also, majority of the contaminated tubes of both media were those inoculated with the morning specimen.
This may be due to the fact that the participants were not directly supervised by the field staff during the production of the morning specimens which could result in contamination with food debris in the mouth unlike the spot one whereby cleaned water was used to rinse the mouth of participants before sputum production. However, due to mitigating measures put in place by laboratory staff with support from the survey coordinating team, the contamination rates declined steadily (13% -17% for MGIT and 5% -9% for LJ) till the end of the survey period. These rates were still higher than the general recommendation of 5% for all media types although up to 7% -8% may be accepted for liquid media [18].

Strengths and Limitations
We highlight here some of the strengths of the laboratory that contributed to the success of the survey. Firstly, the use of two specimens (both spot and morning) as well as two methods for both microscopy and culture accounted for the availability of over 90% of results. Next, specimens were readily processed anytime they arrived from the field due to a 24-hour shift system the laboratory staff run. Thus, except in very rare cases there were little delays for inoculation from time of sputum collection (i.e. within 5 days). Furthermore, the use of fully electronic data collection and entry barcode systems ensured minimal human error. By this, some health facilities have shifted from the fully manual system of data collection and management to semi-automated systems.
Generally, through the survey, the capacity to conduct laboratory examinations such as microscopy (ZN/AO), culture (MGIT/LJ) and GeneXpert have been enhanced. Despite these strengths, some challenges were also encountered. Key among these were the relatively high contamination rates of specimens in the initial stages of the survey which could have underestimated the TB prevalence. Furthermore, due to the high numbers and frequency of specimens received in the laboratory, there could be some possible cases of cross-contamination. However, during the analysis these challenges were treated by imputation analysis. Another notable challenge was the sudden breakdown of the biosafety cabinet in which the samples were being processed at one of the two laboratories. This necessitated temporary suspension of laboratory work for maintenance, and thus delayed inoculation of some of the collected specimens around that period.

Implication for TB Control by NTP
The experiences gained by the laboratory staff have contributed immensely to the improvement of some activities of the NTP. Among these are installation of GeneXpert and hands on training at some regional and district hospitals by some laboratory staff who were involved in the prevalence survey.

Conclusion and Recommendation
In this study, most of the specimens were smear negative but culture positive.
Also, significant number of smear positive specimens was found to be NTM which indicates that relying on smear microscopy alone may lead to misdiagnosis and unnecessary treatment leading to development of drug resistance. Based on these findings and from the laboratory staff point of view we recommend that a wide roll out of novel technologies such as GeneXpert be implemented nationwide. This may increase case detection and minimize false-positive smear and false-negative culture results. Finally, the capacity and skills of the laboratory staff must be enhanced through regular in-service training and other refresher courses.