A total 3675 participants were enrolled at seven study sites over the period June to August 2018 with an overall response rate of 91% (lowest of 79% in Viwandani an urban slum in Nairobi, Kenya and 100% in Ikire and Ogane-Uge both rural areas in Nigeria). Of these, we excluded 109 participants who had no blood pressure measurements and 17 who were missing weight and height measurements. The sample analyzed constituted 3549 participants with a mean age of 39·7 years (SD 15·4) out of which 60·5% were women. Participants in Nigerian sites on average were older than those from East African sites (p < 0·0001). (Table 1)
Table 1
Baseline characteristics, SevenCEWA study 2018
Characteristic | All (n = 3549) | Ikire, Nigeria (n = 489) | Ogane-Uge, Nigeria (n = 403) | Okpok Ikpa, Nigeria (n = 465) | Olorunda Abaa, Nigeria (n = 708) | Soroti, Uganda (n = 760) | Ukonga, Tanzania (n = 424) | Viwandani, Kenya (n = 300) |
Setting | | Semi-urban | Rural | Rural | Rural | Urban | Semi-urban | Urban |
Demographics | | | | | | | | |
Age (years), mean (SD) | 39·7 (15·4) | 48·1 (18·1) | 39·2 (19·5) | 38·5 (14·0) | 41·2 (12·9) | 33·7 (12·4) | 43·9 (14·1) | 34·7 (10·9) |
Women, n (%) | 2147 (60·5) | 269 (55·0) | 211 (52·4) | 225 (48·4) | 458 (64·7) | 534 (70·3) | 305 (71·9) | 145 (48·3) |
Asset index | | | | | | | | |
Poorest | 646 (18·2) | 100 (20·5) | 88 (21·8) | 99 (21·3) | 135 (19·1) | 154 (20·3) | 89 (20·9) | 63 (21·0) |
Poor | 614 (17·3) | 98 (20·0) | 81 (20·2) | 97 (20·9) | 130 (18·4) | 153 (20·1) | 80 (18·8) | 57 (19·0) |
Fair | 742 (20·9) | 98 (20·0) | 166 (41·2) | 131 (28·1) | 135 (19·1) | 152 (20·0) | 82 (19·3) | 60 (20·0) |
Rich | 507 (14·3) | 95 (19·5) | - | 58 (12·5) | 135 (19·1) | 154 (20·3) | 87 (20·5) | 60 (20·0) |
Richest | 575 (16·2) | 98 (20·0) | 53 (13·1) | 80 (17·2) | 133 (18·7) | 147 (19·3) | 86 (20·28) | 60 (20·0) |
Highest level of Education attained | | | | | | | | |
None | 456 (12·8) | 102 (20·9) | 105 (26·1) | 105 (22·6) | 46 (6·5) | 72 (9·5) | 22 (5·2) | 4 (1·3) |
Primary | 1217 (34·3) | 115 (23·5) | 153 (38·0) | 207 (44·5) | 154 (21·8) | 198 (26·1) | 270 (63·7) | 120 (40·0) |
Secondary | 1279 (36·0) | 193 (39·5) | 123 (30·5) | 117 (25·2) | 301 (42·5) | 284 (37·4) | 105 (24·8) | 156 (52·0) |
Tertiary | 580 (16·3) | 68 (13·9) | 22 (5·5) | 32 (6·9) | 205 (29·0) | 206 (27·1) | 27 (6·4) | 20 (6·7) |
Employment status | | | | | | | | |
Self-employed | 2121 (59·8) | 395 (80·8) | 307 (76·2) | 309 (66·5) | 560 (79·1) | 222 (29·2) | 255 (60·1) | 73 (24·3) |
Government employee | 302 (8·5) | 21 (4·3) | 8 (2·0) | 26 (5·6) | 63 (8·9) | 91 (12·0) | 28 (6·6) | 65 (21·7) |
Private employer | 333 (9·4) | 31 (6·3) | 23 (5·7) | 38 (8·2) | 47 (6·6) | 56 (7·4) | 13 (3·1) | 125 (41·7) |
Unemployed | 785 (22·1) | 42 (8·6) | 65 (16·1) | 84 (18·1) | 38 (5·4) | 391 (51·4) | 128 (30·2) | 37 (12·3) |
Refused to answer | 8 (0·2) | - | - | 8 (1·7) | - | - | - | - |
Smoking, n (%) | | | | | | | | |
Never | 3036 (85·5) | 439 (89·8) | 221 (54·8) | 389 (83·7) | 672 (94·9) | 718 (94·5) | 375 (88·4) | 222 (74·0) |
Ever | 409 (11·5) | 47 (9·6) | 114 (28·3) | 71 (15·3) | 31 (4·4) | 42 (5·5) | 26 (6·1) | 78 (26·0) |
Declined to answer | 104 (2·9) | 3 (0·6) | 68 (16·9) | 5 (1·1) | 5 (0·7) | - | 23 (5·4) | - |
Alcohol use | | | | | | | | |
Never | 2690 (75·8) | 427 (87·3) | 226 (56·1) | 256 (55·1) | 672 (94·9) | 633 (83·3) | 349 (82·3) | 127 (42·3) |
Ever | 727 (20·5) | 44 (9·0) | 90 (22·3) | 187 (40·2) | 31 (4·4) | 127 (16·7) | 75 (17·7) | 173 (57·7) |
Declined to answer | 132 (3·7) | 18 (3·7) | 87 (21·6) | 22 (4·7) | 5 (0·7) | - | - | - |
Self-reported Diabetes mellitus, n (%) | 86 (2·4) | 25 (5·1) | 10 (2·5) | 1 (0·2) | 20 (2·8) | 8 (1·1) | 19 (4·5) | 3 (1·0) |
Measurements | | | | | | | | |
Body mass index category, n (%) | | | | | | | | |
Underweight (< 18·5 kg/m2) | 349 (9·8) | 54 (11·0) | 91 (22·6) | 2 (0·4) | 56 (7·9) | 106 (13·9) | 23 (5·4) | 17 (5·7) |
Normal (18·5 to < 25 kg/m2) | 1918 (54·0) | 264 (54·0) | 194 (48·1) | 347 (74·6) | 362 (51·1) | 449 (59·1) | 133 (31·4) | 169 (56·3) |
Overweight/Obese (> 25 kg/m2) | 1159 (32·7) | 139 (28·4) | 110 (27·3) | 110 (23·7) | 250 (35·3) | 181 (23·8) | 256 (60·4) | 113 (37·7) |
Missing | 123 (3·5) | 32 (6·5) | 8 (2·0) | 6 (1·3) | 40 (5·6) | 24 (3·2) | 12 (2·8) | 1 (0·3) |
Waist circumference (both genders) | | | | | | | | |
Men (≥ 102 cm)* | 101 (2·8) | 16 (3·3) | 8 (2·0) | 5 (1·1) | 16 (2·3) | 14 (1·8) | 41 (9·7) | 1 (0·3) |
Women (≥ 88 cm)* | 865 (24·4) | 104 (21·3) | 48 (11·9) | 127 (27·3) | 211 (29·8) | 156 (20·5) | 180 (42·5) | 39 (13·0) |
Blood pressure | | | | | | | | |
Mean SBP (mmHg), mean (SD)^ | 122·9(20·8) | 128·5 (23·9) | 126·2 (20·5) | 126·4 (19·6) | 119·1 (21·6) | 122·8 (17·2) | 125·3 (21·3) | 110·0 (16·3) |
Mean DBP (mmHg), mean (SD)^ | 77·9 (12·8) | 82·0 (14·0) | 77·3 (12·9) | 80·3 (12·4) | 73·8 (12·6) | 80·0 (11·2) | 77·2 (12·7) | 73·4 (11·4) |
Out of the total sample, the sites contributed as follows: 13·8% from Ikire, 11·4% from Ogane-Uge, 13·1% from Okpok Ikpa, 19·9% from Olorunda Abaa, 21·4% from Soroti, 11·9% from Ukonga, and 8·5% from Viwandani. A total 55 participants declined to respond to asset ownership questions 15 (3·7%) in Ogane-Uge and 40 (5·6%) in Olorunda Abaa, Nigeria. |
Across sites, 44% of participants lived in rural areas of Ogane-Uge (11·4%), Okpok Ikpa (13·1%), and Olorunda Abaa (19·9%) all in Nigeria, a quarter lived in semi-urban areas [Ikire, Nigeria (13·8%) and Ukonga, Tanzania (11·9%)], and 29·9% lived in urban communities in Soroti, Uganda (21·4%) and Viwandani, Kenya (8·5%). Participants in Ukonga, Tanzania had the highest prevalence of obesity (60%) whereas those in Ogane-Uge, Nigeria had the highest prevalence of underweight (22%). (Table 1)
Overall, 25·1% (95% Confidence Interval 23·7%, 26·6%) of participants had hypertension. Nigerian communities had the highest crude prevalence of hypertension i.e., 38·6% (34·2%, 43·0%) in Ikire, 33·0% (28·4%, 37·7%) in Ogane-Uge, 23·3% (20·3%, 26·6%) in Olorunda Abaa, and 20·4% (17·9%, 25·6%) in Okpok Ikpa. Among the three East African sites, Ukonga in Tanzania had the highest crude prevalence at 28·5% (24·3%, 33·1%) followed by Soroti in Uganda with 20·4% (17·6%, 23·4%) and the lowest crude prevalence was recorded in Viwandani in Kenya with a 9·7% (6·6%, 13·6%).
The age-standardized prevalence of hypertension was 16·3% (14·5, 18·1) for women and 15·6% (13·5, 17·6) for men. When stratified by site, the age-standardized prevalence was highest in Ogane-Uge, Nigeria at 22·1% (18·0, 26·1) and lowest in Viwandani, Kenya at 11·3% (7·4, 15·1).
Among the 896 participants with hypertension, 43·1% (39·8%, 46·4%) were not aware that they had hypertension. Of those who were known to have hypertension, 47·3% (42·9%, 51·7%) were not taking medications, and of those taking medication 51·6% (44·4%, 56·7%) did not have their blood pressure controlled.
Despite the low prevalence of hypertension in Viwandani (Kenya), about three-quarters [75·9% (56·5%, 89·7%)] of those with elevated blood pressures were not aware that they had hypertension. On the contrary, Nigerian study sites with higher prevalence of hypertension had comparatively higher proportions of awareness of hypertension compared with sites in Tanzania and Kenya.(Fig. 1)
Compared with participants of other sites, participants from Soroti, Uganda and Okpok Ikpa, Nigeria had higher rates of diagnosed but untreated hypertension; 78·9% (70·3%, 86·0%) and 70·5% (60·3%, 79·4%) of those diagnosed, respectively. Overall study sites in Nigeria had higher blood pressure control rates compared to those in east Africa.(Fig. 1)
In models adjusted by site, for both gender, the factors associated with higher mean systolic blood pressures were older age and being overweight/obese. In contrast, being privately employed (compared with unemployed) among both genders, and among women having attained any education (compared with no education) were associated with lower mean systolic blood pressure.(Table 2) Surprisingly, among men current smoking compared to never smoking was associated with a lower mean systolic blood pressure (-4·2 mmHg, 95% CI -7·5, -0·9).
In multivariable analyses, each 10-year increase in age for both sexes was associated with higher odds of prevalent hypertension (adjusted Odds Ratio 1·4, 95%CI 1·4, 1·5), whereas attainment of any education (versus no education) and having health insurance (aOR 0·6, 95%CI 0·5, 0·8) were associated with lower prevalence of hypertension particularly among women.(Table 3) Older age was also associated with a higher odds of awareness for both sexes (aOR 1·2, 95%CI 1·1, 1·3) and primary education was associated with lower odds of awareness among women (aOR 0·5, 95%CI 0·3, 0·7).(Table 3) Finally, older age was also associated with higher odds of treatment for both sexes (aOR 1·2, 95%CI 1·1, 1·3). Having health insurance was also associated with a higher chance of being treated among women (aOR 1·5, 95%CI 1·2, 1·9).(Table 3)
Table 3
Associations of hypertension prevalence, awareness, and treatment for all participants and by gender (adjusted odds ratios and 95% confidence interval), SevenCEWA study 2018
| Both genders | Women | Men |
Prevalence | Awareness | Treatment | Prevalence | Awareness | Treatment | Prevalence | Awareness | Treatment |
Number of participants (number with outcome) (%) | 901 (25·4) | 515 (57·2) | 260 (50·5) | 551 (25·7) | 343 (62·3) | 160 (46·6) | 350 (25·0) | 172 (49·1) | 100 (58·1) |
Age (each 10 yrs) | 1·4 (1·4, 1·5) | 1·2 (1·1, 1·3) | 1·2 (1·1, 1·3) | 1·4 (1·3, 1·5) | 1·2 (1·1, 1·3) | 1·2 (1·0, 1·4) | 1·5 (1·3, 1·6) | 1·2 (1·1, 1·4) | 1·1 (0·9, 1·3) |
Level of Education | | | | | | | | | |
None | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref |
Primary | 0·6 (0·5, 0·8) | 0·5 (0·3, 0·7) | 0·6 (0·4, 0·9) | 0·6 (0·4, 0·7) | 0·4 (0·3, 0·6) | 0·5 (0·3, 0·8) | 0·7 (0·4, 1·2) | 0·5 (0·3, 1·1) | 0·9 (0·4, 2·3) |
Secondary | 0·6 (0·4, 0·8) | 0·7 (0·5, 1·0) | 1·0 (0·6, 1·6) | 0·6 (0·4, 0·9) | 0·7 (0·4, 1·1) | 0·9 (0·5, 1·7) | 0·6 (0·4, 1·1) | 0·6 (0·3, 1·3) | 1·2 (0·5, 3·0) |
Tertiary | 0·7 (0·5, 0·9) | 0·8 (0·5, 1·2) | 1·0 (0·6, 1·8) | 0·6 (0·4, 0·8) | 0·8 (0·5, 1·4) | 1·5 (0·7, 3·1) | 0·9 (0·5, 1·6) | 0·7 (0·3, 1·6) | 0·8 (0·3, 2·2) |
Wealth index | | | | | | | | | |
Poorest | 0·8 (0·7, 1·1) | 0·7 (0·5, 1·0) | 1·1 (0·7, 1·7) | 0·9 (0·7, 1·3) | 0·8 (0·5, 1·2) | 1·1 (0·6, 1·9) | 0·7 (0·5, 1·1) | 0·6 (0·3, 1·1) | 1·0 (0·5, 2·2) |
Poor | 0·9 (0·7, 1·1) | 0·9 (0·6, 1·2) | 0·8 (0·5, 1·4) | 1·0 (0·7, 1·3) | 0·9 (0·6, 1·4) | 0·6 (0·3, 1·1) | 0·7 (0·5, 1·0) | 0·8 (0·5, 1·5) | 1·7 (0·8, 3·8) |
Fair | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref |
Rich | 0·8 (0·7, 1·1) | 0·8 (0·6, 1·2) | 0·9 (0·6, 1·5) | 0·7 (0·5, 1·0) | 0·8 (0·5, 1·3) | 0·7 (0·3, 1·3) | 1·0 (0·7, 1·4) | 0·8 (0·4, 1·4) | 1·5 (0·7, 3·1) |
Richest | 0·9 (0·7, 1·2) | 0·9 (0·6 1·3) | 1·0 (0·6, 1·6) | 0·8 (0·5, 1·1) | 0·8 (0·5, 1·4) | 0·8 (0·4, 1·6) | 1·1 (0·8, 1·6) | 0·9 (0·5, 1·5) | 1·3 (0·6, 2·8) |
Health Insurance | | | | | | | | | |
Uninsured | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref | Ref |
Insured | 0·6 (0·5, 0·8) | 0·8 (0·6, 1·1) | 1·6 (1·4, 1·9) | 0·7 (0·5, 0·9) | 0·9 (0·5, 1·4) | 1·5 (1·2, 1·9) | 0·6 (0·4, 0·8) | 0·7 (0·4, 1·2) | 0·8 (0·4, 1·6) |
*Adjustment was done for study site and gender. |
Although we had smaller numbers and larger uncertainties for analysis of controlled blood pressure as outcome, each decade increase in age was associated with lower odds of control (aOR 0·7, 95%CI 0·6 to 0·8). (Appendix Table 1)
For participants aged 40 years and older, the overall median predicted 10-yr CVD risk of a first fatal and non-fatal CVD (stroke and ischemic heart disease) across all sites was fairly low at 4·9% IQR (2·4%, 10·3%) i.e., for men median 6·5% (IQR 3·7%, 13·1%) and women 3·9% (IQR 1·9%, 8·9%). We excluded men in Okpok Ikpa site because only 7 men aged > 40yrs were enrolled which would give unstable estimates. Noteworthy, the 10-year risk of CVD varied substantially across sites with highest risks estimated in Ikire, Nigeria for both men (median 10·3%, IQR 4·5%, 29·3%) and women (median 9·0%, IQR 5·6%, 33·6%). The lowest predicted 10-yr CVD risk for both gender were in Viwandani, Kenya for men (median 4·7%, IQR 2·6%, 7·4%) and women (median 1·2%, 0·8%, 1·6%). (Fig. 2)
Overall thirteen percent (13·2%) had predicted 10-yr CVD risk of 20% or greater as per the WHO guidelines[18] and 7·1% had predicted 10 year CVD risk using the 30% as the threshold of the global NCD target.[19] (Appendix Fig. 1)