Delay in Diagnosis of Pulmonary Tuberculosis is Associated with Increased Risk of Transmission in Pastoralist Setting, Ethiopia


 Background

To comprehend the effect of delayed care on risk of tuberculosis (TB) transmission in a TB prevalent but low case detection area, this study examined the association of diagnosis delay with patient infectiousness (cavitation and smear positivity) and determined the threshold delay that optimizes infectiousness. It also assessed transmission drivers in Somali region of Ethiopia, an area with ample pastoralist population.

Methods

A cross-sectional study was conducted using 434 new pulmonary TB patients, aged ≥15 years, who were recruited prospectively in five major facilities between December 2017 and October 2018. Data were collected on delays in diagnosis, socio-demographics, clinical and epidemiological information using interview, record-review, anthropometry, sputum microscopy and chest radiography techniques. Log-binomial regression models were used to reveal predictors of cavitation and smear positivity at p<0.05 using Stata/SE®14. C-statistics was applied to determine predictive ability and threshold delay that classifies infectiousness.

Results

Median age of participants was 30 years. Majorities were male (62.9%), nearly half (46.5%) were pastoralist and 2.3% TB/HIV co-infected. Median delay from debut of illness to diagnosis was 49 days (IQR=37). Among all cases, 45.6% [95%CI: 40.9-50.4] had pulmonary cavity and 42.0% [95%CI: 37.3˗46.9] were smear positive. On multivariable analysis, cavitation was higher in patients delayed over a month [P<0.001], ≤35 years [APR (95%CI) =1.3(1.01-1.6)], with chronic diseases [APR (95%CI) =1.8(1.2-2.6)] and low MUAC*female [APR (95%CI) =1.8(1.2-2.8)]. Smear positivity was higher in patients delayed >49 days [p=0.02], ≤35 years [APR (95%CI) =1.4(1.1-1.8)], low BMI [APR (95%CI) =1.3(1.01-1.7)] and low MUAC [APR (95%CI) =1.5(1.2-1.9)]. Delay discriminates cavitation [AUC (95%CI) =0.67(0.62-0.72)] at 43 days optimal cutoff and 74.6% sensitivity.

Conclusion

This study highlights that delay in diagnosis of pulmonary TB remains high and is associated with increased risk of cavitation and smear positivity in pastoral setting in Ethiopia. In pastoral settings, this may call upon a socio-cultural tailored TB prevention and control strategies.


Abstract Background
To comprehend the effect of delayed care on risk of tuberculosis (TB) transmission in a TB prevalent but low case detection area, this study examined the association of diagnosis delay with patient infectiousness (cavitation and smear positivity) and determined the threshold delay that optimizes infectiousness. It also assessed transmission drivers in Somali region of Ethiopia, an area with ample pastoralist population.
Methods A cross-sectional study was conducted using 434 new pulmonary TB patients, aged ≥15 years, who were recruited prospectively in five major facilities between December 2017 and October 2018. Data were collected on delays in diagnosis, socio-demographics, clinical and epidemiological information using interview, record-review, anthropometry, sputum microscopy and chest radiography techniques. Log-binomial regression models were used to reveal predictors of cavitation and smear positivity at p<0.05 using Stata/SE®14. C-statistics was applied to determine predictive ability and threshold delay that classifies infectiousness.

Results
Median age of participants was 30 years. Majorities were male (62.9%), nearly half (46.5%) were pastoralist and 2.3% TB/HIV co-infected. Median delay from debut of illness to diagnosis was 49 days (IQR=37). Among all cases, 45.6% 95%CI: 40.9-50.4 had pulmonary cavity and 42.0% 95%CI: 37.3˗46.9 were smear positive. On multivariable analysis, cavitation was higher in patients delayed over a month P<0.001, ≤35 years APR . The prevalence was found to be higher in pastoral communities (316/100,000) than the national level (277/100,000) [ 2 ]. To curb the global epidemics of this deadliest disease, the End-TB strategy sets early diagnosis and treatment of cases as pillars to ending TB epidemics by 2030 [ 3 ]. For the most part, the national TB control program (NTP) of Ethiopia detects TB cases when people with presumptive symptoms present themselves to health facilities (passive case finding strategy) [ 4 ].
However, this passive approach struggles to achieve the required case detection rates in resource-limited settings, allowing millions of potentially infectious cases undiagnosed in communities [ 5 , 6 ]. Nearly one-third of TB cases in Ethiopia were not notified in 2017 [ 1 ]. Local-specific reports indicated the number of undetected infectious cases in community equals the number of notified cases [ 7 ], and another study reported up to two-thirds of active cases remained undetected by the passive system [ 6 ]. In Somali Regional State of Ethiopia (SRS), new case detection rate has not exceeded 50% in recent years [ 8 ]. This implies a high number of infectious cases exist in households and communities without obtaining proper diagnosis and treatments, which is likely influenced by patients' healthcare seeking behavior and health system deficiencies [ 9 , 10 ]. Extreme delay in diagnosis and treatment of TB has been challenging in Ethiopia, and median delay exceeding two months was reported in pastoral settings [ 11 ].
Failures to timely detect cases and initiate treatment worsen the disease, increase risk of death, increase risk of treatment failure and drug resistance, and exacerbate ongoing transmission in households and congregate settings. Devastating damages occur in lung tissues as patients delay longer without proper treatment, the classical hallmark is cavity formation [ 12 , 13 ]. Cavities are sites of excessive TB bacilli accumulation and release higher bacilli load in aerosols [ 14 ]. Moreover, cavities slow smear conversion following treatment (prolongs contagious period), and are associated with high treatment failure and relapse, emergence of drug resistance, disease dissemination and permanent lung impairments [ 15 ]. Findings have indicated a call for extended treatment of Cavitary TB with a combination of new drugs and new treatment strategies [ 16 ], yet no special strategy is currently in place in Ethiopia.
In addition to influencing infectiousness, delay prolongs contagious period and extends contact time between index case and close contacts [ 17 ]. This highlights the need for assessing the effect of delays on risk of TB transmission to assess the effectiveness of TB control programs in controlling the disease and interrupting its transmission. However, the effect of delays on risk of transmission, to the best of our knowledge, was not addressed in pastoral settings in Ethiopia. There is also limited data on cavitary TB and the acceptable delays from clinical and programmatic perspectives. Hence, this study was intended to assess the association of delay in diagnosis with infectiousness of patients and determine threshold delays that optimize cavitation and smear positivity as proxy measures of infectiousness. We have also assessed household drivers of transmission in Somali regional state of Ethiopia where majorities of the population lead pastoral life and households inhabit in cramped transitory huts [ 18 ].

Study Setting
Four hospitals (Kharamara, Dege-habour, Kebri-Daher and Gode) and one health center (Abilelie) in Somali Regional State of Ethiopia were selected purposefully based on their patient flow, presence of radiologic facility, and geographic location in the administration.
Kharamara hospital and Abilelie health center are located in the regional capital, Jigjiga.
The rest facilities are found in less-urbanized, pastoral-dominant and semi-arid zones of the region. Approximately 85% of the region's population lead a nomadic or agro-pastoral way of life [ 18 ]. The nomads rear livestock, migrate seasonally while agro-pastoralists are relatively permanent, and carry out mixed herding and farming [ 11 ]. The hospitals and selected health centers provide TB services as per the National guideline, which involves two spot-spot smear microscopy examination spaced by   4 ]. People in this age category are believed to acquire competent immunity that is key in cavity formation [ 19 ], cover 80% of all TB cases and account for almost 100% of disease transmissions [ 20 ]. Patients with lung co-morbidities (bronchitis, pneumonia and lung cyst) were excluded.

Sample Size and Sampling Technique
The minimum sample size estimated using OpenEpi303 for cross-sectional studies was 282. This assumed 95% CI, 80% power, 1:1 ratio of non-delayed/delayed, 27.5% of nondelayed and 45% of delayed patients had cavitation in related study [ 12 ], 5% precision and 10% non-response rate, and given delay above 30 days as critical point at which risk of transmission increases [ 21 ]. We included all the available samples in the analysis to increase the power of the test, which raised the final sample size to 434. Patients were recruited sequentially from the first date of data collection. As the patients arrived to the Directly Observed Therapy-Short Course (DOTS) facilities for treatment initiation, all upcoming eligible PTB patients were recruited for the study before initiating treatment.
Data Collection: Questionnaire, Microscopy and Chest X-ray A mix of methods including interview, anthropometry, Acid-Fast Bacilli (AFB) microscopy and chest radiography were used in addition to the standard medical examination (record review). A structured and pre-tested questionnaire was employed to obtain data on delays in diagnosis, socio-demographics, self-reported medical conditions, and environmental drivers of transmission. Records were reviewed to substantiate co-morbidities. Mid-Upper Arm Circumference (MUAC) was measured using inelastic paper tapes, and Body Mass Index (BMI) was computed from weight (kilograms) and height (meter-square) measures.
Nurses working in DOTS clinics carried out recruitment, interview, record review and anthropometry procedures. Training was provided on sampling and data collection procedures by the principal investigator and a local research assistant.

AFB Examination
Upon completion of interviews, the DOTS providers linked patients to radiology and laboratory units using request forms prepared for this purpose. Three sputum specimens from each patient were collected; morning sputum at home, and two spot specimens spaced by 30 minutes after the patient delivered the morning specimen. A pair of smears was prepared from each specimen, air dried and heat fixed. One slide of each pair was examined at hospital laboratories using Ziehl Neelsen (  All patients underwent Chest X-ray examinations to identify lung cavitation, measure cavity size and count the number of cavities. A senior radiologist at Kharamara hospital examined all the X-ray films and digital imaging. The radiologist was blinded to radiologic and AFB results reported during the standard initial diagnosis. Sample of X-ray films (n=41) were randomly picked and blindly re-checked by another radiologist to ensure the reliability of X-ray readings. As of rechecking, we found levels of 95.1% [84. 6

Data Processing and Analysis
Data were double entered and validated using EpiData version 3.1; and analyzed using Stata/SE ® 14 (StataCorp, College Station, Texas 77845 USA). Descriptive statistics was performed to summarize delays in diagnosis, patient infectiousness, explanatory and environmental factors of transmission. Prevalence ratios along with 95% confidence intervals (CI) were used to compare cavitation and smear positivity between categories of predictors, and multivariable analyses were fitted using Log-binomial regression models.
Statistical significances were determined at p-value ≤0.05; and p-value ≤ 0.2 in bivariate analysis was used as a cutoff point for inclusion in final models. C-statistics or Receiver Operating Characteristic (ROC) was employed to determine the discriminatory ability and threshold/optimal cutoff points of diagnosis delay that classify patient infectiousness at maximum sum of sensitivity and specificity, and positive likelihood ratio (LR+), given sensitivity (>70%).

Operational/standard definition of terms
Pulmonary Tuberculosis: is a patient with lung TB of either smear-positive or negative forms. A smear AFB positive patient is confirmed if at least one AFB positive smears; A smear negative patient is diagnosed if: at least two AFB smear negative results, no response to a course of broad-spectrum antibiotics, again two AFB negative smears and radiological abnormalities consistent with TB; Or two AFB smear negative results but culture positive for MTB [ 4 ].
New Case: is a patient who has never had treatment for TB before or has not yet initiated anti-TB treatment.
Retreatment case: is a patient who was treated for any form of TB before but has developed the disease again following relapse or default or failure to cure during the 1 st regimen.
Newly Diagnosed Patient: a patient who was prospectively diagnosed with Pulmonary TB during the study period. This excludes patients who were on treatment.
Diagnosis delay: is defined as the period from debut of the first symptom(s) particularly cough or other (chest pain, haemoptysis, weight loss, night sweating) to the date of TB diagnosis.
Infectiousness: is the capability of a PTB patient to transmit TB infection into a susceptible person, characterized by the existence of pulmonary cavity and/or AFB positive smear.
Pulmonary cavity: is an air-containing lucent space within a consolidation or a mass or nodule surrounded by infiltrate or fibrotic wall identified upon radiological examination [ 19 , 22 ].
The median diagnosis delay from debut of respiratory symptoms to the date of TB diagnosis was 49 days (IQR=37), ranging 8 to 362 days. Four rural patients received care longer than 254 days after the onset of the early respiratory illnesses.

Discussion
The present finding reveals that close to half (45.6%) of all and 82.5% of smear positive patients with pulmonary TB had one or more cavities; 42% were sputum smear positive, and half of them delayed more than seven weeks and few nearly a year without medical care. Cavitation increased continuously as patients delayed longer than four weeks, and optimized at threshold delay of 43 days. Similarly, smear positivity was higher in patients who delayed above seven weeks.
Cavitation and smear positivity were reciprocally illustrative, and the majority of patients had either cavitation or was smear positive. Cavities are the stockpiles of mycobacterial accumulation, and connected to airflow they release high bacillary load in sputum and nasal droplets, the channel for transmission [ 23 ]. This connotes the large majority of patients had intricate form of the disease and was capable of transmitting TB prior to diagnosis or treatment. This cavitation rate matches with the maximum assumption of 50% rate that happens if patients do not Cavitation and smear positivity were notably higher in delayed patients. The median delay was higher than the threshold delay (43 days) that optimizes the risk of cavitation, and it was the significant delay at which smear positivity increases. To be precise, the majority of patients (60%) delayed above the threshold delay of cavitation without obtaining care.

Availability of Data and Materials
The dataset supporting the conclusions of this article is included within the article. The collected data contain confidential information, and consent has not been obtained for public sharing of raw data with identifiers. However, the datasets used and/or analyzed are available at the hands of the corresponding author and can be shared upon reasonable requests.        Figure 1 Map of the study area Box plot illustrating the distribution of diagnosis delay in days Area under the ROC curve of diagnosis delay as a prognosis test of pulmonary cavitation Figure 6 The predicted probability of pulmonary cavitation at each value of the observed diagnosis delay Figure 7 Area under the ROC curve of diagnosis delay as a prognosis test of smear positivity