Chronic respiratory disease in adult outpatients in three African countries: a cross-sectional study

BACKGROUND : The greatest burden of chronic respiratory disease is in low- and middle-income countries, with recent population-based studies reporting substantial levels of obstructive and restrictive lung function. OBJECTIVE : To characterise the common chronic respiratory diseases encountered in hospital outpatient clinics in three African countries. METHODS : This was a cross-sectional study of consecutive adult patients with chronic respiratory symptoms (>8 weeks) attending hospital outpatient departments in Ethiopia, Kenya and Sudan. Patients were assessed using a respiratory questionnaire, spirometry and chest radiography. The diagnoses of the reviewing clinicians were ascertained. RESULT: A total of 519 patients (209 Kenya, 170 Ethiopia, 140 Sudan) participated; the mean age was 45.2 years (SD 16.2); 53% were women, 83% had never smoked. Reviewing clinicians considered that 36% (95% CI 32–40) of patients had asthma, 25% (95% CI 21–29) had chronic bronchitis, 8% (95% CI 6–11) chronic obstructive pulmonary disease (COPD), 5% (95% CI 4–8) bronchiectasis and 4% (95% CI 3–6) post-TB lung disease. Spirometry consistent with COPD was present in 35% (95% CI 30–39). Restriction was evident in 38% (95% CI 33–43). There was evidence of sub-optimal diagnosis of asthma and COPD. CONCLUSION : In Ethiopia, Kenya and Sudan, asthma, COPD and chronic bronchitis account for the majority of diagnoses in non-TB patients with chronic respiratory symptoms. The suboptimal diagnosis of these conditions will require the widespread use of spirometry.

The greatest burden of chronic respiratory disease is in low-and middle-income countries, with recent population-based studies reporting substantial levels of obstructive and restrictive lung function. O B J E C T I V E : To characterise the common chronic respiratory diseases encountered in hospital outpatient clinics in three African countries. M E T H O D S : This was a cross-sectional study of consecutive adult patients with chronic respiratory symptoms (.8 weeks) attending hospital outpatient departments in Ethiopia, Kenya and Sudan. Patients were assessed using a respiratory questionnaire, spirometry and chest radiography. The diagnoses of the reviewing clinicians were ascertained. R E S U LT: A total of 519 patients (209 Kenya, 170 Ethiopia, 140 Sudan) participated; the mean age was 45.2 years (SD 16.2); 53% were women, 83% had never smoked. Reviewing clinicians considered that 36% (95% CI 32-40) of patients had asthma, 25% (95% CI [21][22][23][24][25][26][27][28][29] had chronic bronchitis, 8% (95% CI 6-11) chronic obstructive pulmonary disease (COPD), 5% (95% CI 4-8) bronchiectasis and 4% (95% CI 3-6) post-TB lung disease. Spirometry consistent with COPD was present in 35% (95% CI [30][31][32][33][34][35][36][37][38][39]. Restriction was evident in 38% (95% CI [33][34][35][36][37][38][39][40][41][42][43]. There was evidence of sub-optimal diagnosis of asthma and COPD. C O N C L U S I O N : In Ethiopia, Kenya and Sudan, asthma, COPD and chronic bronchitis account for the majority of diagnoses in non-TB patients with chronic respiratory symptoms. The suboptimal diagnosis of these conditions will require the widespread use of spirometry. K E Y W O R D S : chronic respiratory symptoms; spirometry; COPD; asthma; Africa; Ethiopia; Kenya; Sudan; hospital clinics Chronic respiratory diseases (CRDs) are leading noncommunicable diseases worldwide, most probably because of the ubiquity of poverty, and inhaled noxious environmental, occupational and behavioural exposures. 1 CRDs affect more than 1 billion people worldwide, with asthma and chronic obstructive pulmonary disease (COPD) being the most prevalent: about 300 million people have asthma and 200 million have COPD. [1][2][3] The greatest burden of CRDs occurs in low-and middle-income countries (LMICs): almost 90% of COPD deaths and 80% of asthma deaths occur in LMICs, where they have been linked with poverty, poor access to healthcare and limited health service resources. 4 Capacity for CRD diagnosis and management, such as spirometry and inhaled therapies, are not generally available in primary health care in public health services. [5][6][7][8] Among the 1.1 billion people who live in sub-Saharan Africa, population-based surveys report wide variations in asthma symptom prevalence (5.7-20.3%), being highest in 'westernised' urban settings. 9 The Burden of Obstructive Lung Disease (BOLD) study reports that the prevalence of fixed airflow obstruction in sub-Saharan African countries ranges from 9-73% in men and 4-86% in women. 10 BOLD and other studies also report variations in the prevalence of restricted spirometry (4-49%). [11][12][13] The aim of this clinic-based study was to describe and estimate the prevalence of common CRDs in patients with chronic respiratory symptoms attending hospital outpatient departments in three sub-Saharan African countries.  14 Consecutive adult patients aged 18 years with chronic respiratory symptoms (.8 weeks) 15 reviewed in the participating hospital outpatient departments were invited to take part. Exclusion criteria wee clinical suspicion of TB, positive GeneXpert sputum test result, pregnancy, acute respiratory infection or contraindications for spirometry.

Data collection
An interviewer administered a respiratory questionnaire based on that used in BOLD with additional questions developed as part of the National Institute for Health Research International Multidisciplinary Programme to Address Lung Health and TB in Africa ''IMPALA''. The questions covered demographics, symptoms, medical history, exposures to outdoor and indoor pollutants, tobacco smoking and occupation. 16 The diagnosis made by the reviewing clinician was ascertained.
Pre-and post-bronchodilator (BD) spirometry was performed as detailed below. A postero-anterior chest radiograph was also performed and systematically read/recorded by a radiologist. 17 Spirometry Spirometry was conducted in accordance with Pan African Thoracic Society (PATS), American Thoracic Society/European Respiratory Society recommendations by trained and PATS-certified technicians/ nurses in Competence in Foundational Spirometry. 18,19 The EasyOne Spirometer (NDD Medizintechnik, Zurich, Switzerland) was used with daily calibration checks. Up to eight forced exhalation manoeuvres were performed while sitting. All traces were reviewed by an external assessor, and measurements graded A-C for acceptability and repeatability were selected for analysis. The spirometric parameters recorded were forced expiratory volume in 1 sec (FEV 1 ), forced vital capacity (FVC) and the ratio (FEV 1 /FVC). Spirometry measurements were compared with 2012 Global Lung Initiative reference equations for 'Black' ethnicity. 20 Patients' lung function was categorised as follows: normal pre-BD FEV 1 /FVC LLN (lower limit of normal), FEV 1 LLN and FVC LLN; fixed-airflow obstruction (COPD) post-BD FEV 1 /FVC , LLN; and restrictive pattern FVC , LLN. Significant BD reversibility was defined as an improvement in FEV 1 12% of baseline and 200 ml. 21 Two patterns of spirometry were considered consistent with asthma: 1) significant reversibility with pre-BD airflow obstruction, FEV 1 /FVC , LLN (strict classification), 2) significant reversibility alone (looser classification). 22 As per GOLD (Global Initiative for Chronic Obstructive Lung Disease) guidelines, post-BD airflow obstruction (FEV 1 /FVC , 0.7) was categorised into mild (FEV 1 80% predicted), moderate (80% . FEV 1 50% predicted), severe (50% . FEV 1 30% predicted) and very severe (FEV 1 , 30% predicted). 21

Statistical considerations
Descriptive data are presented as percentages or mean, with standard deviation (SD) and 95% confidence intervals (CIs), where appropriate. Two operational definitions of asthma prevalence are presented 'clinician diagnosed' and 'current asthma' (wheeze in the last year). 22,23 Between-group comparisons were performed using v 2 , Fisher's exact test, t-tests or analysis of variance, as appropriate; P , 0.05 was considered statistically significant. We aimed to recruit 200 participants from each country (total sample size of 600), this would enable the study to estimate disease prevalence with 2% precision assuming disease prevalence was 5-10%. Analyses were performed using IBM SPSS Statistics for Windows, v25.0 (IBM, Armonk, NY, USA).

RESULTS
In total, 519 patients (209 Kenya, 170 Ethiopia, 140 Sudan) participated; the numbers completing the different study assessments are presented in Supplementary Table S1. All completed the questionnaire; acceptable pre-and post-BD spirometry was available for respectively 448 and 447 patients. Overall and individual country participant demographic characteristics are presented in Table 1. Participants tended to be female (53%) and middle-aged (mean age: 45 years, SD 16), with a minority having ever smoked (17%). About 15% of patients lived in rural environments, and nearly a third (29%) reported that they had gone short of food in the previous year. HIV seropositivity was reported by 4.1% and a previous history of TB by 18%.
Cough, sputum expectoration and breathlessness were associated with reduced FEV 1 , FVC, and evidence of obstruction and restriction (Supplementary Table S3). Breathlessness was associated with significant reversibility. Pure restriction was not associated with an excess of symptoms. Cliniciandiagnosed asthma was associated with increased wheeze but not cough, sputum expectoration or breathlessness (Table 4). Overall, clinician-diagnosed asthma was not associated with any differences in FEV 1 , FVC, or FEV 1 /FVC, but it was associated with BD reversibility and pure obstructive pattern of spirometry (25% vs. 14%). 'Current asthma' (i.e., wheeze in last year) was associated with reduced FEV 1 , FVC, FEV 1 /FVC, BD reversibility and pure obstructive and mixed patterns of spirometry (Supplementary Table S3). Cliniciandiagnosed COPD was associated with an increase in the symptoms of sputum expectoration and breathlessness but not cough or wheeze (Table 4). Clinician-diagnosed COPD was not associated with any significant differences in FEV 1 , FVC, FEV 1 / FVC; however, the number of people with diagnosed COPD was small (n ¼ 41, 7.9%).

Lung function and diagnosis of airways disease
Given the lack of associations between clinician diagnoses of asthma or COPD and FEV 1 , and obstruction, further analyses were undertaken (Table  5). Of the 159 patients with clinician-diagnosed asthma and satisfactory spirometry, 62 (39%) fulfilled the spirometric criteria for COPD. Of the 267 patients without clinician-diagnosed asthma and satisfactory spirometry, about 20% had spirometry consistent with asthma: 47 (18%) had evidence of airflow obstruction and significant reversibility and 55 (21%) had significant reversibility alone (looser classification). Of the 36 patients with a previous COPD diagnosis and satisfactory spirometry, 23 (64%) did not fulfil the spirometric criteria for COPD (Table 6). Of the 390 without clinician-diagnosed COPD who provided satisfactory spirometry, 139 (36%) fulfilled the spirometric criteria for COPD, which is considerably more than those with clinician-diagnosed COPD.

DISCUSSION
This cross-sectional study was conducted in general hospitals of three African countries to estimate the prevalence of common CRDs in patients presenting to outpatient clinics with chronic respiratory symptoms. Consistent with the global CRD burden, the most common diseases reported by patients enrolled in this study were asthma, COPD and chronic bronchitis, accounting for about 70% of patients. 24 The prevalence of clinician-diagnosed asthma and  * 'Asthma þ reversibility þ AFO' is a sub-group of 'Asthma þ reversibility'. AFO ¼ airflow obstruction, i.e., post-bronchodilator FEV 1 /FVC,LLN; FEV 1 ¼ forced expiratory volume in 1 sec; FVC ¼ forced vital capacity; LLN ¼ lower limit of normal.
'current asthma' (wheeze in the last year) were respectively 36% (95% CI 32-40) and 71% (95% CI 66-74). Reviewing clinicians diagnosed chronic bronchitis in 25% (95% CI [21][22][23][24][25][26][27][28][29], and although reviewing clinicians diagnosed COPD in 7.9% (95% CI 5.7-10.6) of patients, spirometry consistent with COPD was present in 35% (95% CI 31-40). A further aim was to determine, using spirometry, whether the substantial levels of restrictive lung function reported in community surveys in sub-Saharan Africa would be evident in everyday clinical practice. Although 38% (95% CI 33-43) of the patients had evidence of restrictive lung function, in nearly 60% there was concomitant fixed airflow obstruction. Purely restrictive spirometry was present in 17% (95% CI 14-21) of patients and not associated with an excess of symptoms. Depending on strictness of spirometric classification, between 27% and 33% of patients with clinician-diagnosed asthma had spirometry consistent with asthma; however, it is recognised that objective measures of airflow obstruction correlate poorly with asthma symptoms and response to treatment. 25 We also found that 39% of clinician-diagnosed asthma patients fulfilled the criteria for COPD, and while this may be a consequence of chronic airway remodelling, this finding raises the probability of COPD being misdiagnosed as asthma, a well-recognised issue when diagnoses are totally reliant on symptoms unsupported by spirometry. 26 In addition to asthma overdiagnosis, there was also evidence for the converse, with 18-21% (depending on strictness of spirometric classification) of symptomatic patients with no asthma diagnosis having spirometry consistent with asthma; numerically, such patients accounted for more than those with spirometry supportive of their clinician-diagnosed asthma. This is consistent with studies reporting underdiagnosis of asthma if the diagnosis is unsupported by spirometry testing. [26][27][28] Clinician-diagnosed COPD was reported in 8% of patients; however, only 36% had evidence of fixed airflow obstruction. To note, 36% of symptomatic patients with no COPD diagnosis fulfilled the spirometric criteria for COPD, 45% had been diagnosed with asthma by the reviewing clinician and 47% had severe/very severe airflow obstruction, a proportion consistent with the BOLD study findings. 29 Patients with undiagnosed COPD (n ¼ 139) considerably exceeded those with diagnosed COPD (n ¼ 41). This finding in symptomatic patient populations adds to the very limited data on the under-diagnosis of COPD in Africa, 30 and almost certainly reflects the lack of routine access to spirometry.
In Kenya, the most common abnormal pattern of lung function was obstruction, whereas in Ethiopia and Sudan there was a greater prevalence of restriction than in Kenya, with a predominance of a mixed obstruction/restriction pattern. Pure restriction was also more common in Ethiopia and Sudan than Kenya. When compared with Kenya, in addition to having higher rates of diagnosed asthma or COPD, Ethiopia and Sudan also appeared to have higher rates of undiagnosed asthma or COPD. These differences are likely to reflect between-country differences in influential factors accumulated through the life course, e.g., prematurity, birth weight, infant weight gain, diet, air pollution, smoking, occupation(s), TB.
The findings of our study differ from those of a previous study of patients attending clinics in Khartoum, Sudan, that reported 3% of patients had pure obstruction, 6% mixed obstruction/restriction and 5% pure restriction; however, this study was conducted in patients with previous smear-positive TB. 31 Our findings from Kenya are similar to those of 762 patients referred to the respiratory units of a teaching hospital in Ghana, which had a predominance of obstructive lung disease, with 48% of patients having normal spirometry, 26% pure obstruction, 15% pure restriction and 12% mixed obstruction/restriction. 32 In a Ugandan study, 17% of 792 patients attending a hospital chest clinic were reported to have diagnosed asthma; however, no spirometry was available. 33 We found that BD reversibility was more prevalent in Ethiopia (35%) than in Kenya (24%), which is consistent with the Table 6 Patterns of airflow obstruction in those with and those without clinician-diagnosed COPD findings of the African Severe Asthma Study Project (ASAP) conducted in Ethiopia, Kenya and Uganda. 34 Our study had a number of strengths. Consecutive patients with respiratory symptoms for more than 8 weeks were recruited, the diagnosis of the reviewing clinician was ascertained, and TB was excluded on clinical grounds and a negative GeneXpert test. We were able to conduct high-quality pre-and post-BD spirometry in the majority of patients, giving us insight into COPD and asthma prevalence. Although we did not achieve our recruitment target of 200 in Ethiopia and Sudan, the sample sizes were based on very conservative 5-10% prevalence rates for asthma and COPD; in reality, the prevalence rates for asthma and COPD were three times higher.
Chronic respiratory disease i