CLINICAL RESEARCH

1
Prevalence of CKD, Diabetes, and 52
2 Hypertension in Rural Tanzania 53
3 54
4 Q26 David W. Ploth1, Jessie K. Mbwambo2, Virginia A. Fonner3, Bruce Horowitz4, Phillip Zager5, 55
5 Ron Schrader5, Francis Fredrick6, Caroline Laggis7 and Michael D. Sweat3 56
6 Q2 1
57
Department of Medicine, Division of Nephrology, Medical University of South Carolina, Charleston, South Carolina, USA;
7 2
Department of Psychiatry and Muhimbili University of Health and Allied Sciences, Dar es Salaam, Tanzania; 3Department of 58
8 Psychiatry and Behavioral Sciences, Division of Global and Community Health, Medical University of South Carolina, 59
9
Q3 Charleston, South Carolina, USA; 4Division of Nephrology, University of Utah School of Medicine, Salt Lake City, Utah, USA;
5
60
Department of Medicine, Division of Nephrology, University of New Mexico, Albuquerque, New Mexico, USA; 6School of
10 Medicine, Muhimbili University of Health and Allied Sciences, Dar es Salaam, Tanzania; and 7Department of Dermatology,
61
11 University of Utah School of Medicine, Salt Lake City, Utah, USA 62
12 63
13 64
Introduction: Chronic kidney disease (CKD), diabetes, and hypertension play a disproportionate role in the
14 65
growing public health challenge posed by noncommunicable diseases (NCDs) in East Africa. The impact of
15 these NCDs may pose the greatest challenge in rural areas with limited screening and treatment facilities, 66
16 although precise prevalence estimates of these conditions in rural Tanzania are lacking. 67
17 Methods: The prevalence of CKD, diabetes, and hypertension, were estimated from a probability sample 68
18 of adults (n ¼ 739) residing in 2 communities within Kisarawe, a rural district of Tanzania. Following 69
19 consent, participants were studied in their homes. Random point-of-care (POC) measures of glycosylated 70
20 hemoglobin and blood pressure, were obtained. Serum creatinine, drawn at the POC and measured at 71
Muhimbili National University, was used to calculate estimated glomerular filtration rate with the Chronic 72
21 Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.
22 73
Results: The median age was 35 years (interquartile range 2545 years). Overall the pooled prevalence for 74
23 Q4 CKD stages 3, 4, and 5 was 12.4% (95% confidence interval [CI] ¼ 10.214.8). Surprisingly, the prevalence
24 of CKD stage 5 (3.0%; 95% CI ¼ 2.14.4) was high among the youngest age group (1836 years). The 75
25 prevalence estimates for prehypertension and hypertension were 38.0% (95% CI ¼ 34.641.5) and 19.9% 76
26 (95% CI ¼ 17.122.9), respectively. The prevalence estimates for prediabetes and diabetes were 25.7% 77
27 (95% CI ¼ 22.629.1) and 14.8% (95% CI ¼ 12.417.6), respectively. 78
28 Conclusion: Although this pilot study had a relatively small sample size, the prevalence estimates for CKD, 79
29 diabetes, and hypertension were higher than we expected based on previous estimates from Tanzania. 80
30 CKD was not significantly associated with diabetes or hypertension, suggesting an alternative etiology. 81
31 Kidney Int Rep (2018) -, -–-; https://doi.org/10.1016/j.ekir.2018.04.006 82
32 KEYWORDS: diabetes mellitus; hypertension; kidney disease; prevalence; Tanzania 83
ª 2018 International Society of Nephrology. Published by Elsevier Inc. This is an open access article under the CC BY-
33 NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
84
34 85
35 86
36 CDs are the leading cause of death worldwide.1 Relatively few studies have been conducted on the 87
37
38
N Prominent among NCDs is the worldwide
epidemic of CKD and the related risk factors of diabetes
prevalence of NCDs, including CKD, diabetes, and
hypertension, in East Africa.5 Stanifer et al. reported
88
89
39 and hypertension.2 The impact of this epidemic is that the prevalence of CKD was higher in the Moshi 90
40 especially severe in low- and middle-income countries urban (15.2%; 95% CI ¼ 9.6%23.3%) versus the 91
41 with limited health facilities for screening and treat- Moshi rural (7.0%; 95% CI ¼ 3.8%12.5%) district in 92
42 Q5 ment.2–4 Obtaining precise estimates of the prevalence the Kilimanjaro Region of Tanzania.5 Janmohamed et al. 93
43 of CKD, diabetes, and hypertension in areas with reported that among a group of adult patients in a 94
44 limited health resources is the first step in decreasing diabetes mellitus clinic of Bugando Medical Centre in 95
45 the burden imposed by these NCDs. Mwanza, Tanzania, the prevalence of CKD (detected by 96
46 the presence of proteinuria) was 83.7%.6 97
47 Correspondence: David W. Ploth, Division of Nephrology, The Therefore, the purpose of this study was to evaluate 98
48 Medical University of South Carolina, 96 Jonathan Lucas Street, the prevalence of CKD, diabetes mellitus (DM), and 99
49 MSC 629, CSB, Charleston, South Carolina 29425, USA. E-mail:
hypertension (HTN) in a random sample of participants Q6 100
plothdw@musc.edu
50 with POC measures in rural Tanzania. The present 101
Received 4 December 2017; revised 29 March 2018; accepted 16
51 April 2018; published online 22 April 2018 study supplemented a National Institutes of Health 102

Kidney International Reports (2018) -, -–- 1

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce

 CLINICAL RESEARCH DW Ploth et al.: Chronic Kidney Disease, Diabetes, and Hypertension Q1

103 (NIH)/National Institute of Mental Health (NIMH) selected participants within households. In addition, 157
104 funded study on HIV prevention strategies (Compre- community outreach was performed prior to data 158
105 hensive Triaged HIV Prevention Trial) in 2 commu- collection to raise awareness about the study and to 159
106 nities in the Kisarawe District, Tanzania. We examined familiarize community members with study staff. 160
107 the prevalence of CKD, DM, and HTN in the same Eligible participants had to be 18 to 55 years of age, 161
108 communities. Our overarching goal was to develop a to reside in the household, and to plan on continuing to 162
109 better understanding of the prevalence of these dis- reside in the community for the next 2 years. The 163
110 eases and to understand the contributing factors so that assessment for each participant included completing a 164
111 we can develop effective strategies to blunt the long- behavioral survey with a trained research assistant and 165
112 term health consequences for these areas with limited completing a biological assessment that included POC 166
113 health care resources. The study was approved by the testing for HIV and other sexually transmitted diseases, 167
114 Institutional Review Boards at the Medical University as well as blood pressure (BP) measurement and POC 168
115 of South Carolina, Charleston, South Carolina, USA, and HbA1c testing. A venous blood sample was also drawn 169
116 Muhimbili University of Health and Allied Sciences, for transport to a laboratory for measurement of serum 170
117 Dar es Salaam, Tanzania. creatinine. Behavioral survey data and biological data 171
118 were captured using an electronic tablet (Samsung Tab 172
119 METHODS 2) and uplinked via cellular telephone to the central 173
120 The present study was designed as a supplement to a data repository-data storage server. Written informed 174
121 funded phase II community-based randomized consent was provided by each participant prior to data 175
122 controlled trial assessing HIV prevention strategies in collection. The consent document was read aloud to all 176
123 rural Tanzania (Triage Project). The parent trial took participants to ensure comprehension regardless of 177
124 place within 2 rural communities, with approximately literacy level. 178
125 10,000 inhabitants each, in Kisarawe District, Tanzania. 179
126 Data for the present study were collected during the Measures 180
127 baseline assessment of the Triage Project among a Behavioral Survey 181
128 community cohort (n ¼ 739 individuals, 372 partici- The behavioral survey was administered in Kiswahili 182
129 pants from community A and 367 participants from by trained interviewers in or near the participants’ 183
130 community B). Kisarawe is a rural district located homes. The survey mostly pertained to sociodemo- 184
131 approximately 40 km from Dar es Salaam and is part of graphic information and sexual risk behavior, although 185
132 the Pwani (Coast) Region of Tanzania. participants were asked whether they had ever been 186
133 diagnosed with DM, HTN, or CKD and, if so, whether 187
134 Sampling Scheme they had received treatment. 188
135 Subjects in the community cohort were randomly 189
Blood Pressure Measurement
136 sampled using a 2-stage sampling scheme. All house- 190
Participants’ BP was measured in the household with a
137 hold locations within the 2 communities were 191
LifeSource UA-767PlusAC Blood Pressure Monitor
138 enumerated and mapped using Global Positioning Sat- 192
(A&D Medical/LifeSource, Mississauga, ON, Canada).
139 Q7 ellite (GPS) coordinates. Household locations were 193
Blood pressure was measured according to American
140 randomly selected from the list of all 3714 enumerated 194
Heart Association (AHA) guidelines,7 with 3 measure-
141 household locations in the 2 communities (1920 195
ments obtained and averaged for a final value. Hy-
142 household locations in community A and 1794 in 196
pertension was classified using American Heart
143 community B). A household was defined as family 197
Association criteria for pre-HT and HTN: BP 120139/
144 members (whether related or not related by blood) who 198
8089 mm Hg; stage 1 HTN 140159/9099 mm Hg,
145 live on the same compound and eat from the same pot. 199
and stage 2 HTN $160/$ 100 mm Hg.7
146 Second, all eligible household members of selected 200
147 households were identified and enumerated. One HbA1c 201
148 participant was selected from each household contain- HbA1c was measured at POC with a Bayer A1c Now kit 202
149 ing eligible members using a random selection appli- (Bayer Healthcare AG, Leverkusen, BRG). Blood sam- 203
150 cation programmed into a Samsung Tab 2 Tablet ples for the measurement of HbA1c were obtained by 204
151 Q8 Computer. Participants were sampled without replace- fingerstick. Both pre-DM (HbA1c $ 5.7% to # 6.4%) 205
152 ment at the household level. The randomly identified and DM (HbA1c $ 6.5%) were diagnosed in accor- 206
153 participant was recruited to participate in both the dance with American Diabetes Association criteria8 and 207
154 parent HIV study and simultaneously contribute to our the World Health Organization.9 To validate the HbA1c 208
155 pilot NCD study. To reduce nonresponse rates, up to 3 Now test in this setting, we compared the results for 209
156 attempts were made to contact selected households and determination of HbA1c levels with the POC A1c Now 210
2 Kidney International Reports (2018) -, -–-

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce

 DW Ploth et al.: Chronic Kidney Disease, Diabetes, and Hypertension CLINICAL RESEARCH

211 test and a laboratory-based test (using SIEMENS DCA and stage 2 HTN) was used to ascertain probability of 265
212 Q9 Analyzers) among 20 randomly selected blood speci- group membership based on age, gender, and an age- 266
213 mens from the Muhimbili Hospital blood bank. We by-gender interaction. For DM, ordinal logistic 267
214 analyzed results by calculating the mean difference regression was used to assess the relationship of DM 268
215 (% A1c) between the 2 tests and used a paired t test to stage (normal, prediabetic, and diabetic) with age, 269
216 assess the significance of the differences between the gender, and an age-by-gender interaction. The rela- 270
217 results between the 2 tests. tionship of eGFR with comorbid DM or HTN were 271
218 Serum Creatinine assessed using 1-way analysis of variance, grouping by 272
219 A venous blood sample was obtained and transported stage of DM or HTN. A Welch 2-sample t test was used 273
220 refrigerated (2  C to 8  C) to Muhimbili National Uni- to compare eGFR levels when both HTN and DM were 274
221 versity Hospital, for measurement of serum creatinine present versus not having the 2 comorbid conditions. 275
222 concentration. Transit time was <60 hours. After 276
223 arrival at the laboratory, samples were frozen (20  C) 277
224 RESULTS 278
until analyzed. Serum creatinine was measured by an
225 Q10 IDMS standardized alkaline picric acid method.10 Demographic and medical information data acquired 279
226 Serum creatinine concentrations were used to calcu- from the questionnaire for the NCD pilot study are 280
227 late estimated glomerular filtration rate (eGFR) using shown in Table 1. Of 1237 households visited by study 281
228 the CKD-EPI equation.11 CKD was classified according staff, nonresponse rates by household across commu- 282
229 to U.S. National Kidney Foundation Kidney Disease nities were similar (101 of 569 in community A and 124 283
230 Outcomes Quality Initiative (KDOQI) criteria and the of 668 in community B), but community B had signif- 284
231 National Kidney Foundation Guidelines for eGFRs <60 icantly more households containing no eligible house- 285
232 ml/min per 1.73 m2.12,13 hold members (n ¼ 147) than did community A (n ¼ 286
233 88). Of the remaining 777 participating households 287
234 Statistical Analyses with eligible members, 38 selected participants either 288
235 Data were analyzed with R (Vienna, Austria: Founda- refused to participate or could not be located, with 289
236 tion for Statistical Computing, Vienna, Austria)14 and significantly more refusals occurring in community B 290
237 Stata v14 (StataCorp LLC, College Station, TX).15 Dif- (n ¼ 30) than in community A (n ¼ 8). The overall 291
238 ferences between demographic characteristics between nonresponse rate was 4.8% (38 of 776). A total of 739 292
239 the 2 communities were measured using a c2 statistic participants provided data (372 in community A and 293
240 for categorical variables and a t test for continuous 367 in community B) (Figure 7). Although 739 partici- Q27 294
241 variables. For continuous variables, the median value pants gave informed consent and contributed behav- 295
242 and interquartile range were reported. Prevalence of ioral questionnaire information and demographic data 296
243 HTN, CKD, and DM were calculated as proportions for the study, measurements of serum creatinine were 297
244 using the number of participants screened for each not available for 26 participants. Similarly, data for 298
245 disease as the total population at risk, and 95% CIs HbA1C were not available for 31 participants, and BP 299
246 were calculated for each proportion. Prevalence esti- measurements were not available for 4 participants. 300
247 mates were weighted by the inverse of the probability The median age was 35 years (interquartile range 301
248 of being selected at the household level to account for 2545 years) with similar distribution across the 2 302
249 variations in household size and by community strati- communities. The sample population was 59.3% female 303
250 fication, accounting for the 2 communities’ unequal (n ¼ 438), which differed from 2012 census data 304
251 sizes. Regarding CKD, the relationship of eGFR to age reporting a 51% female population nationally. 305
252 and gender was descriptively examined using locally Farming was the predominant occupation. Nearly all 306
253 weighted scatterplot smoothing (LOESS), and analysis of the population were self-employed, compatible with 307
254 of covariance was used to detect overall differences in a subsistence lifestyle that reflects the economy of the 308
255 eGFR by age and gender. The eGFR was categorized study area. Importantly, 9.6% of the population 309
256 into 5 groups according to CKD stage, and multinomial completing the questionnaire (738 responses) indicated 310
257 logistic regression was conducted to assess the proba- knowledge of pre-existing HTN, and 1.6% (739 re- 311
258 bility of group membership using age, gender, and the sponses) had knowledge of pre-existing DM. Approx- 312
259 interaction of age and gender as independent variables. imately one-half of those reporting a history of either 313
260 Regarding diabetes and hypertension, LOESS smooth- HTN or DM also reported having been on treatment for 314
261 ing was used to assess the overall relationship of the respective disorder. Current therapy was not 315
262 HbA1c and BP (systolic and diastolic) with age and questioned. Approximately 1% of those responding 316
263 gender. For HTN, ordinal logistic regression (ordered (737 responses) related knowledge of pre-existing kid- 317
264 groups included: normal BP, pre-HTN, stage 1 HTN, ney disease, and about one-third of those individuals 318
Kidney International Reports (2018) -, -–- 3

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce

 CLINICAL RESEARCH DW Ploth et al.: Chronic Kidney Disease, Diabetes, and Hypertension

319 Table 1. Demographic and medical data for study population 373
320 Community A (N [ 372) Community B (N [ 367) Overall (N [ 739) P value Q20 374
321 Q21 Age, median (IQR) 35 (2745) 36 (2844) 35 (2545) 0.87 375
322 Gender 0.20 376
323 Male 160 (43.0%) 141 (38.42%) 301 (40.7%) 377
324 Female 212 (57.0%) 226 (61.58%) 438 (59.3%)
378
Ethnicity <0.001
325 Zaramo 196 (52.7%) 293 (79.8%) 489 (66.2%)
379
326 Other 176 (47.3%) 74 (20.2%) 250 (33.8%) 380
327 Occupation 0.001 381
328 Self-employed farmer 264 (70.97%) 215 (58.80%) 479 (64.8%) 382
329 Self-employed vendor 54 (14.52%) 91 (24.80%) 145 (19.6%) 383
330 Not employed, seeking employment 21 (5.65%) 18 (4.90%) 39 (5.3%)
384
Employed with a fixed salary 19 (5.11%) 15 (4.09%) 34 (4.6%)
331 Other or not seeking employment 14 (3.76%) 28 (7.63%) 42 (5.7%)
385
332 Education 0.04 386
333 No formal education 81 (21.77%) 54 (14.71%) 135 (18.3%) 387
334 Primary education 250 (67.20%) 271 (73.84%) 521 (70.5%) 388
335 High school and secondary education 41 (11.02%) 42 (11.44%) 83 (11.2%) 389
336 Monthly income (USD), median (IQR) 13.64 (0.4545.45) 18.18 (4.5545.45) 18.18 (2.2745.45) 0.08
390
337 Data are numbers with percentages in parentheses unless otherwise indicated. IQR, interquartile range; USD, U.S. dollars. 391
338 392
339 reported some treatment for kidney problems. Based on found to have CKD stage 4, 0.8% (95% CI ¼ 0.31.9) 393
340 the questionnaire responses, disease awareness was and 24 individuals CKD stage 5, 3.0% (95% CI ¼ 394
341 extremely low. 2.14.4). Estimated GFR decreased with advancing age 395
342 The overall estimated prevalence of CKD stages 3 for the population as a whole (generalized logistic 396
343 through 5 was 12.4% (95% CI ¼ 10.214.8) in 712 regression, P # 0.05) (Figure 1). This effect was more 397
344 individuals for whom creatinine measurements were pronounced for females than for males. Furthermore, 398
345 available (Table 2). There were 97 of 712 individuals eGFR was observed to be lower with advanced age for 399
346 with eGFR <60 ml/min per 1.73 m2 (Table 2). Stage 3 individuals with CKD (generalized logistic regression, 400
347 disease was the most prevalent level of CKD in this P # 0.05). Examined by individual stage of CKD, 401
348 population, with 68 of 712 individuals with eGFR advancing age was significantly associated with 402
349 reported between 30 and 59 ml/min per 1.73 m2, 8.5% increased frequency of sCKD stage 3 (n ¼ 64; P < 0.05). 403
350 (95% CI ¼ 6.810.6). In contrast, 5 individuals were However, this pattern was not significant for CKD stage 404
351 4 or 5, perhaps because of the smaller sample sizes (n ¼ 405
352 5 and n ¼ 24, respectively). Moreover, there was a 406
353 Q22 Table 2. Prevalence and characteristics of chronic kidney disease surprisingly high frequency of CKD stage 5 observed in 407
354 in the study population the youngest age group. Of 24 individuals with CKD 408
355 n Casesa Prevalence estimate (95% CI)b stage 5, 37.5% (n ¼ 6 of 24) were <26 years of age. 409
356 Q23 Overall CKD, stages 35 712 97 12.4 (10.214.8) Although there was no significant age-by-gender 410
357 Gender interaction, females exhibited an increased proportion 411
358 Male 292 41 13.5 (10.217.7)
of eGFR between 30 and 90 ml/min per 1.73 m2 and a 412
Female 420 56 11.5 (8.914.7)
359 Age, yr
decreased proportion of eGFR >90 ml/min per 1.73 m2 413
360 1830 151 18 11.7 (7.517.7) relative to males (P < 0.05) (Fig. 2a and b). Estimated 414
361 3139 202 21 10.2 (6.715.3) CKD prevalence differed significantly by community. 415
362 4049 212 30 13.0 (9.417.9) The estimated prevalence of CKD in community A was 416
363 $50 147 28 15.4 (10.821.3) 20.8% (95 CI ¼ 16.925.2) and 4.4% (95% CI ¼ 417
364 Community
2.86.9) in community B. 418
Community A 358 80 20.8 (16.925.2)
365 Community B 354 17 4.41 (2.86.9)
Among the 735 individuals for whom BP measure- 419
366 Comorbidities ments were available, the estimated prevalence of stage 420
367 Hypertension 118 14 12.3 (10.115.0) 1 or stage 2 HTN was 19.9% (95% CI ¼ 17.122.9) 421
368 Diabetes 105 15 12.1 (7.518.8) (Table 3). The pre-HTN estimated prevalence was 422
369 Self-reported kidney problems, n (%) 739 6 0.8% 38.0% (95% CI ¼ 34.641.5), stage 1 HTN was 15.6 423
370 Treated for kidney problems 6 2 33%
(95% CI ¼ 13.018.4), and stage 2 HTN was 4.3% 424
371 CI, confidence interval; CKD, chronic kidney disease.
a
Cases refer to cases of CKD (stages 35).
(95% CI ¼ 3.15.9). Similar to the association of 425
372 b
Weighted to account for household size and community stratification. increasing age with lower eGFR, logistic regression 426
4 Kidney International Reports (2018) -, -–-

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce

 DW Ploth et al.: Chronic Kidney Disease, Diabetes, and Hypertension CLINICAL RESEARCH

427 Table 3. Prevalence and characteristics of hypertension in the Q24 481

428 study population 482
429 n Casesa Prevalence (95% CI)b 483
430 Overall HTN, stages 1 and 2 734 144 19.9 (17.122.9) 484
431 Gender 485
432 Male 299 49 17.0 (13.121.7)
486
Female 435 95 21.9 (18.226.1)
433 Age, yr
487
434 1830 158 16 12.3 (7.718.9) 488
435 3139 212 32 15.0 (10.820.5) 489
web 4C=FPO

436 4049 215 44 23.6 (18.230.0) 490

437 $50 149 52 31.7 (25.039.3) 491
438 Community
492
Community A 369 69 20.0 (16.024.6)
439 Community B 365 76 19.7 (16.124.0)
493
440 Figure 1. Relationship of estimated glomerular filtration rate (eGFR) Self-reported HTN, n (%) 739 71 (9.6%) 494
441 to age. Scatterplot of eGFR versus age with genders identified as Treated for HTN 71 38 495
female (red) and male (black). Locally weighted scatterplot
442 smoothing (LOESS) smooths have been fitted for each gender. CI, confidence interval; HTN, hypertension. 496
443 497
a
Cases refer to cases of hypertension (stages 1 and 2).
Analysis of covariance confirmed that eGFR decreases with age b
Weighted to account for household size and community stratification.
444 similarly for both genders. Overall age is significant (P < 0.001), with 498
445 on average a 12eGFR unit drop per decade of life. Genders differ 499
446 significantly (P < 0.001), with females on average 8.7 eGFR units and stage 2 HTN. Correlation of eGFR with stage of 500
lower than males. There is no significant gender-by-age interaction HTN failed to reveal any significant relationship
447 501
(P ¼ 0.9). (KruskalWallis test, P ¼ 0.86) (Figure 4). We also
448 502
449 could not detect a significant relationship with analysis 503
450 showed significantly higher systolic or diastolic BP of covariance for the presence of HTN when the stage 504
451 with advancing age for the study population as a whole of HTN was correlated with the level of GFR after 505
452 (P < 0.05) (Figure 3a and b). Proportional odds logistic adjusting for the effect of age (P > 0.1). 506
453 regression analysis demonstrated significantly greater An estimated 25.7% of the population (95% CI ¼ 507
454 increases in BP with age in females as compared with 22.629.1) exhibited prediabetic levels of HbA1c, 508
455 males (P < 0.001). This was evident for both stage 1 whereas 14.8% (95% CI ¼ 12.417.6) were estimated 509
456 510
457 511
458 512
459 513
460 514
461 515
462 516
463 517
464 518
465 519
466 520
467 521
web 4C=FPO

468 522
469 523
470 524
471 525
Figure 2. Relationship of (a) estimated glomerular filtration rate (eGFR) group by age and (b) eGFR group by age and gender. (a) Chronic kidney
472 disease (CKD) stage categorized according to eGFR < 15, 15 # eGFR < 30, 30 # eGFR < 60, 60 # eGFR < 90, and 90 # eGFR. Multinomial
526
473 logistic regression (unordered categories or generalized logistic regression) depicts the estimated probability of group membership for each of 527
474 the 5 eGFR groupings at every age; at any given age, the estimated probabilities sum to 1. Because there are so few patients in the bottom 2 528
475 eGFR groups (those with eGFR <30), particularly for older ages, the model does not find any significant trend with age for those groups (P > 0.5 529
476 for both). Proportions between 30 and 60 and between 60 and 90 increase significantly with age, and, as a consequence, proportion >90 530
decreases significantly with age (P < 0.05 for all). (b) Similar to the eGFR value grouping shown in (a), eGFR group are depicted by age and
477 531
gender. Predicted estimated probability curves by gender are identified as female (red) and male (black). The gender-by-age interaction was
478 not significant (P > 0.3), so the overall trends are similar for the genders. However, there is a significant gender effect that shows increased 532
479 proportions for eGFR between 30 and 90 and decreased proportion >90 for females relative to males (that is the result of eGFR in females being 533
480 slightly lower than in males). 534
Kidney International Reports (2018) -, -–- 5

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce

 CLINICAL RESEARCH DW Ploth et al.: Chronic Kidney Disease, Diabetes, and Hypertension

535 589
536 590
537 591
538 592
539 593
540 594
541 595
542 596
543 597
544 598
web 4C=FPO

545 599
546 600
547 601
548 602
Figure 3. Relationship of hypertension to age and gender. Systolic and diastolic blood pressure (BP) are plotted against age, separated by Q19
549 gender, identified as female (red) and male (black). locally weighted scatterplot smoothing (LOESS) smooths have been fitted for each gender.
603
550 Although systolic and diastolic BP increases for both genders with increasing age, the effect for females is more pronounced. Proportional odds 604
551 logistic regression confirmed a significant gender-by-age interaction (P < 0.001). 605
552 606
553 to be overtly diabetic (Table 4). Similar to the associa- HbA1c using the POC test was 5.31% and 5.39% using 607
554 tion of advancing age with increased BP and decreased the laboratory-based test. The mean difference between 608
555 levels of eGFR, levels of HbA1c increased with tests was 0.085% (standard deviation ¼ 0.50%). Re- 609
556 advancing age (Figure 5a). Proportional odds logistic sults from the t test showed no significant difference 610
557 regression analysis (Figure 5b) demonstrated no sig- between the 2 testing strategies (P ¼ 0.46). 611
558 nificant gender-by-age interaction (P > 0.05) with No significant relationship could be detected be- 612
559 HbA1c, although both age (P < 0.001) and gender were tween eGFR or stage of CKD when HTN (stages 1 or 2) 613
560 significantly associated with the stage of diabetes and abnormal HbA1c were treated as a combined, 614
561 (P < 0.002). Whereas average HbA1c levels for females single comorbidity (Welch 2-sample t test, P ¼ 0.23). 615
562 were only slightly greater than those for males, females The extent of association of HbA1c elevation reflecting 616
563 had an increased likelihood of falling in the pre-DM DM, elevated BP reflecting the presence of HTN, and 617
564 and DM groups at all ages (P < 0.002). Also evident CKD are shown in the Venn diagram in Figure 6. 618
565 from Table 4 and Figures 5a and b, the prevalence of 619
566 DM increased with advancing age. When we compared 620
567 results from the POC HbA1c test and the laboratory- DISCUSSION 621
568 based test among the 20 randomly selected specimens Given the increasing burden of NCDs in low- and 622
569 from the Muhimbili Hospital blood bank, the mean middle-income countries, this study provides additional 623
570 624
571 625
Table 4. Prevalence and characteristics of diabetes in the study Q25
572 626
population
573 627
n Casesa Prevalence (95% CI)b
574 628
Overall diabetes (HbA1c $6.5) 707 105 14.8 (12.4L17.6)
575 Gender
629
576 Male 289 38/289 13.1 (9.7L17.6) 630
577 Female 418 67/418 16.0 (12.8L19.9) 631
578 Age, yr 632
579 1830 149 17/149 11.4 (7.2L17.6) 633
580 3139 202 19/202 9.4 (6.1L14.3)
634
4049 210 31/210 14.8 (10.6L20.2)
581 $50 146 38/146 26.0 (19.5L33.8)
635
582 Community 636
583 Community A 356 60 17.4 (13.8L21.7) 637
584 Figure 4. Relationship between stage of hypertension (HTN) and Community B 351 45 12.6 (9.5L16.5) 638
estimated glomerular filtration rate (eGFR). One-way analysis of
585 Self-reported diabetes n (%) 739 12 (1.6%) 639
variance reveals no significant difference between individuals with
586 normal blood pressure (BP) through stage 1 hypertension. However,
Treated for diabetes 12 6
640
587 those with stage 2 hypertension do exhibit significantly lower values
CI, confidence interval.
a
Cases refer to cases of diabetes (HbA1c $ 6.5).
641
588 for eGFR than individuals in the other stages (P < 0.01). b
Weighted to account for household size and community stratification. 642
6 Kidney International Reports (2018) -, -–-

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce

 DW Ploth et al.: Chronic Kidney Disease, Diabetes, and Hypertension CLINICAL RESEARCH

643 697
644 698
645 699
646 700
647 701
648 702
649 703
650 704
651 705
652 706
web 4C=FPO

653 707
654 708
655 709
656 710
Figure 5. Relationship between measured HbA1c, age, and gender. (a) Scatterplot of HbA1c versus age of patient, separated by gender (red
657 711
represents female, black represents male). Locally weighted scatterplot smoothing (LOESS) smooth fits are included. (b) Estimated probabilities
658 of diabetes (DM) stage (normal, pre-DM, DM) by age of patient, separated by gender. Calculated from proportional odds logistic regression. 712
659 There was no significant gender-by-age interaction, but both age (P < 0.001) and gender (P < 0.002) were significant. The probability of being in 713
660 the pre-DM or DM stage increases with age, and females are more likely to be in those stages than are males at all ages. 714
661 715
662 716
663 information regarding the prevalence of CKD, HTN, and greater prevalence of CKD than males at all ages. 717
664 DM among a population-based sample in rural Tanzania. Although no comparative data for age and gender for 718
665 We report data derived from a probability-based sample CKD are available for a more general population in 719
666 of 739 individuals from a rural environment with POD Tanzania, similar trends of decreasing eGFR with age 720
667 measurements obtained in the household. are noted in the U.S. population.17 However, there are 721
668 We observed an overall prevalence of stages 3 to 5 no evident gender differences in the prevalence of CKD 722
669 CKD of 12.4%. In contrast, data reported for the Na- stages 1 to 4 reported for National Health and Nutrition 723
670 tional Health and Nutrition Examination Survey 1999 Examination Survey IV data.16 724
671 to 2004 by the National Center for Health Statistics A recent systematic review and meta-analysis of 725
672 indicated the prevalence of CKD stages 3 to 5 at 5.8% in CKD in sub-Saharan Africa reported the overall CKD 726
673 the United States.16 Therefore, the prevalence of CKD prevalence to be 13.9% (95% CI ¼ 12.2%15.7%), as 727
674 (stage 3 and above) in the population studied in rural synthesized from 21 studies,18 which is similar to 728
675 Tanzania appears to exceed the prevalence of CKD of findings from this study. However, Stanifer et al., in a 729
676 the same stages in the United States. We found that the study conducted in northern Tanzania, found signifi- 730
677 prevalence of CKD in rural Tanzania increases directly cant differences between urban and rural CKD preva- 731
678 with increasing age, and that females exhibited a lence (15.2% and 2.0%, respectively).5 Stanifer et al. 732
679 identified CKD as eGFR <60 ml/min or persistent pro- 733
680 teinuria, and by these criteria the overall community Q11 734
681 Diabetes N = 104 prevalence of CKD reported was 7.0%.5 In an addi- 735
CKD N = 95
682 tional paper, Stanifer et al. reported a significant clus- 736
683 64 11 70 tering of multiple NCDs, including CKD, across urban 737
684 and rural neighborhoods in Tanzania, suggesting that 738
685 4 etiologies may be driven by environmental and life- 739
686 25 10 style factors.19 In the present study, we also found 740
687 disparate prevalence of CKD based on the community 741
688 77
of residence, although both communities were rural. 742
689 More research is needed to understand what is driving 743
Hypertension N = 116
690 the variation in CKD prevalence between neighbor- 744
web 4C=FPO

691 hoods and communities in Tanzania. Based on these 745

692 440 observations, it is clear that CKD poses a significant 746
693 health threat and health care burden in Tanzania, and 747
694 No hypertension, diabetes, CKD N = 440 more understanding of disease etiology is needed. 748
695 Q28 Figure 6. Venn diagram displaying the co-associations of chronic It was surprising that we could not identify the ex- 749
696 kidney disease (CKD), hypertension, and HbA1c. pected association of DM and HTN with CKD in our 750
Kidney International Reports (2018) -, -–- 7

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce

 CLINICAL RESEARCH DW Ploth et al.: Chronic Kidney Disease, Diabetes, and Hypertension

751 Community A Community B 805

752 806
753 1920 Mapped household
807
1794 Mapped household
754 locaons locaons 808
755 809
756 810
757 569 Households visited by Nonresponse (N = 101) 668 Households visited by Nonresponse (N = 124) 811
758 study staff • Not contacted aer 3 visits
study staff
• Not contacted aer 3 visits 812
(N = 76)
759
(N = 45)
• Refusal by head of household • Refusal by head of household 813
760 (N = 56) (N = 48) 814
Ineligible households (N = 147)
761 Ineligible households (N = 88)
• No eligible household • No eligible household members 815
762 380 Eligible parcipants members (N = 88)
397 Eligible parcipants
(N = 147) 816
763 idenfied idenfied 817
764 Refusal/could not be Refusal/could not be 818
web 4C=FPO

765 located aer 3 visits located aer 3 visits 819

766 (N = 8) (N = 30)
820
767 372 Parcipants completed 367 Parcipants completed 821
768 assessment assessment 822
769 Figure 7. Flowchart displaying study sampling, participant recruitment, and nonresponse rates. 823
770 824
771 study. This is in striking contrast to data for the United CKD other than the usual DM and HTN risk factors 825
772 States, where some 85% or greater of CKD is associated prevalent in the United States and Europe. 826
773 with DM and/or HTN.20 However, another study con- The present study is 1 of the few to report BP mea- 827
774 ducted in Tanzania by Stanifer et al. reported that more surements by age and gender for individuals in 828
775 than one-half of CKD cases were not associated with the Tanzania. Overall, results from the present study are 829
776 usual risk factors, including HTN and DM.5 Thus, based consistent with others in the region. For example, a 830
777 on these studies, it is unclear whether the causality of recent study in northern Tanzania found the prevalence 831
778 CKD among Tanzanian populations is related to the of hypertension to be 17.6% (95% CI ¼ 13.921.9) 832
779 traditional risk factors, including HTN and DM, through among rural residents.22 Another study found the 833
780 some more subtle mechanism or the result of some yet-to- prevalence of hypertension in urban Tanzania to be 834
781 be-identified postinfectious, environmental, genetic, or 23.7% (95% CI ¼ 21.3%26.2%).23 Our observations 835
782 other etiology of kidney disease. An additional report indicate that BP significantly increased with age, and at 836
783 from northwestern Tanzania reported an extremely high age >55 years, the BP increase was greater for females 837
784 CKD prevalence of 83.7% among diabetic patients than for males. Another study in northern Tanzania that 838
785 attending a hospital-based DM clinic (CKD defined as assessed age- and gender-stratified prevalence of HTN 839
786 eGFR <60 ml/min per 1.73 m2, or the presence of found that HTN was significantly correlated with 840
787 microalbuminuria or proteinuria).6 The prevalence of increasing age and was nonsignificantly higher among 841
788 CKD in this population was likely biased by the highly males as compared to females.24 Stanifer et al. reported a 842
789 selected nature of a diabetic population being evaluated prevalence of HTN of 30.6% and 19.1% for urban and 843
790 or treated at a tertiary medical facility. In addition, this rural settings, respectively, in Northern Tanzania, 844
791 study did not take into account potential infectious although results were not stratified by gender or age.5 845
792 agents and parasitic agents, such as streptococcus, Data reported for NHANES 1999 to 2000 indicated a 846
793 schistosomiasis, Brugia Malayi (causing filarisis), and prevalence of stage 1 and stage 2 hypertension in 24.6% 847
794 malaria, that are prevalent in the region and could be (95% CI ¼ 22.526.8) in the U.S population.16 Blood 848
795 contributing to the high prevalence of CKD in this pressure increased with advancing age in the United 849
796 geographic area near Lake Victoria.21 States, and, although the BP values are higher for males 850
797 Results from the present study suggest that a sur- at younger ages, females demonstrate significantly 851
798 prising number of young people have CKD stage 5. To greater increases in BP than males after age 55 years, in 852
799 our knowledge, this is the first finding of high CKD concert with our observations for this rural sample in 853
800 prevalence biased toward this younger age group in Tanzania.25 854
801 eastern, sub-Saharan Africa. However, it should be The prevalence of DM in the present study was 855
802 noted that these estimates are based on limited numbers higher than in several other studies in the region. For 856
803 of individuals in these groups. However, if true, this example, a study from northwestern Tanzania and 857
804 trend could represent an unknown causal factor for southern Uganda found that the DM prevalence ranged 858
8 Kidney International Reports (2018) -, -–-

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce

 DW Ploth et al.: Chronic Kidney Disease, Diabetes, and Hypertension CLINICAL RESEARCH

859 from 1% to 2%, as measured through random and CKD etiology. In other parts of the world, including 913
860 fasting blood glucose readings.22 Another study from Central America, Egypt, Sri Lanka, and India, instances 914
861 Tanzania found the adult DM prevalence to be 3.7%, of CKD of unknown etiology are increasing, specifically 915
862 also measured using blood glucose.23 However, a more among males involved in agricultural activities. A recent 916
863 recent study from Tanzania that also used POC HbA1c publication postulated a role for uric acid in the patho- 917
864 test kits found a 21.7% (95% CI ¼ 15.2–29.8) preva- genesis of Mesoamerican nephropathy in Nicaragua, a 918
865 lence of glucose impairment (defined as potential candidate for causality of CKD in this Tanza- 919
866 HbA1c $6%), which is comparable to the present
26
nian population.34 A recent systematic review of CKD of 920
867 study’s findings. It is possible that the use of POC unknown etiology identified some commonalities in risk 921
868 HbA1c tests overestimated DM prevalence. A study factors across regions, including exposure to heavy 922
869 from Uganda compared DM prevalence using 2 metals and agriculture, but the review noted that risk 923
870 methods, namely, POC HbA1c testing and fasting factors were not consistent across several geographic 924
871 plasma glucose, among a population-based sample of areas.35 Evaluation of the causality of CKD of unknown 925
872 more than 700 individuals. The study found that etiology within sub-Saharan Africa merits further study. 926
873 11.3% of the sample had DM when diagnosed using The strengths of the present study include the study 927
874 the HbA1c measure compared to only 4.8% when us- design and POC measures. Households were randomly 928
875 ing fasting plasma glucose.27 HbA1c can be spuriously selected based on random geographic mapping of a 929
876 affected by vitamin and mineral deficiencies, illnesses, defined, statistically relevant sample of a local popu- 930
877 and hematologic factors. Acute or chronic blood loss lation for 2 different areas. An additional strength was 931
878 results in lower HbA1c, although iron deficiency, the POC testing used to obtain all of the measurements 932
879 which might be expected to be common in this largely in the household immediately after obtaining informed 933
880 vegetarian population, results in elevation of HbA1c.28 consent. However, the results of this study also have 934
881 Although intake of carotenoids may decrease, other several limitations. CKD was assessed with a single 935
882 nutritional deficiencies such as Mgþþor chromium in- measurement of serum creatinine concentration in a 936
883 crease HbA1c.29 In addition, hemoglobinopathies such blood sample acquired at the POC but requiring 937
884 as sickle cell could significantly affect HbA1c levels of transport to a distant (although CLIA-certified) labo- Q12 938
885 individuals in this region.30 It is possible that these ratory. Although we have no evidence to suggest 939
886 factors affected the HbA1c levels of participants, irre- inaccurate measures, it is possible that sample degra- 940
887 spective of their DM status, which could have led to an dation during transportation prior to serum separation 941
888 overestimation of prevalence. altered creatinine levels.36 Chronic kidney disease was 942
889 The relatively high apparent prevalence for the identified from estimated GFR derived from the CKD- 943
890 NCDs that were examined in this pilot study was un- EPI equation. The CKD-EPI equation used to estimate 944
891 expected. Although similar estimates of prevalence GFR has not been validated for this population, 945
892 were reported earlier for other areas, we expected them although efforts to evaluate the equation among 946
893 to be relatively lower in this rural, agricultural area of various ethnicities and races, including populations 947
894 the country. Also unexpected to us is that the preva- from sub-Saharan Africa, have been conducted else- 948
895 lence estimates for HTN, DM, and kidney disease that where.37,38 A recent study from Kenya found the 949
896 we observed are similar to what have been reported in CKD-EPI equation to be the most accurate estimating 950
897 high-income countries and specifically in the United equation when comparing results to direct measure of 951
898 States.4,31,32 GFR by iohexol clearance among a small sample of HIV- 952
899 In addition, we found no significant correlation of infected individuals, although the authors also found 953
900 CKD with DM and HTN, which represent 2 of the most that eliminating the race coefficient improved accuracy 954
901 common traditional risk factors in the United States and further.39 We were unable to obtain systematic obser- 955
902 Europe. The recent study on CKD in northern Tanzania vations for the presence of proteinuria or other markers 956
903 conducted by Stanifer et al.5 found that approximately of kidney disease in the present study, which did not 957
904 one-half of CKD cases were not associated with tradi- allow estimation of the prevalence of CKD stages 1 and 958
905 tional DM and HTN risk. Taken together, these findings 2 or more robust estimates for the prevalence of CKD 959
906 suggest likely contributions from infectious, environ- stages 3 to 5. Diabetes was assessed only with HbA1c 960
907 mental, genetic, or other yet-to-be- identified causes. measurement at the POC; we did not measure fasting 961
908 Although potential infectious causes of CKD and its plasma glucose. Although care was taken to maintain Q13 962
909 precursors (e.g., chronic glomerulonephritis) in this re- the appropriate temperature for the HbA1c test kits 963
910 gion have been acknowledged, including hepatitis C, prior to use and although the kits “lock-out” if exposed 964
911 hepatitis B, and HIV,33 no comprehensive study in sub- to extremely high temperature, this could be a variable 965
912 Saharan Africa has been undertaken to fully investigate that was difficult to control or verify. Our validation of 966
Kidney International Reports (2018) -, -–- 9

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce

 CLINICAL RESEARCH DW Ploth et al.: Chronic Kidney Disease, Diabetes, and Hypertension

967 the HbA1c POC test kits comprised a small sample size, for NCDs inadequate in most facilities.41 However, the 1021
968 which limited our ability to detect systematic differ- Tanzanian government is currently working toward 1022
969 ences between the POC and laboratory tests. It is also strengthening the diagnostic and treatment capacity for 1023
970 possible that using an HbA1c cut-off of 6.5% to di- DM and HTN. Furthermore, the focus of most available 1024
971 agnose diabetes is not appropriate for this population, health care is on treatment of acute, infectious condi- 1025
972 but we used this cut-off because it is recommended by tions, especially febrile illnesses. The knowledge base, 1026
973 the World Health Organization and the American technology, and resources, such as pharmaceuticals, 1027
974 Diabetes Association. In addition, we were unable to focused on diagnosis and treatment of these chronic, 1028
975 collect data on height and weight. It is also possible NCDs needs to become a health priority with stronger 1029
976 that there is some survivor bias in this pilot study, in donor support. 1030
977 that patients with the most severe HTN, DM (particu- In conclusion, this study identified a relatively high 1031
978 larly type I DM), and kidney disease may have been prevalence of NCDs including CKD, HTN, and DM 1032
979 lost prior to the data acquisition for our study, which among a rural population in Tanzania. Given the costs 1033
980 could explain the overrepresentation of sCKD stage 5 and numerous challenges involved with treating these 1034
981 Q14 seen among young people. Because we could only NCDs,42 especially in the advanced stages, it is imper- 1035
982 measure HbA1c, we could not separately identify type ative that we develop low-cost, effective prevention 1036
983 1 from type 2 DM in our sample. However, advanced and early treatment paradigms in low-resource settings 1037
984 CKD and severe HTN were present in sufficient in order to reduce the risk of renal failure and the 1038
985 numbers that some early mortality related to severe expected high cardiovascular morbidity associated 1039
986 disease would be expected and could have blunted the with it. 1040
987 prevalence estimates that we report. Whereas HbA1c 1041
988 levels tend to be stable for 2 to 3 months, serum DISCLOSURES 1042
989 creatinine levels can vary over the course of a day, All the authors declared no competing interests. 1043
990 depending on nutrition, hydration, drug interactions, 1044
991 and presence of infections.40 The study population ACKNOWLEDGMENTS 1045
992 comprised community members 18 to 55 years of age; Funding for this project was received from the Global Q15 1046
993 thus elderly patients were excluded, and estimates Health Initiative Pilot Project Award, Center for Global 1047
994 among this population were not obtained. Because of Health, Medical University of South Carolina and Dialysis 1048
995 the age restriction within our sample and the gender Clinics Inc., Nashville, TN, and the National Institute of 1049
996 imbalance, our results are not as generalizable to the Health Award NIMH RO1MH095869. 1050
997 larger Tanzania population, as the sample does not 1051
998 mirror the entire age and gender structure of the REFERENCES 1052
999 population. In addition, whereas our 2-stage sampling 1. GBD 2015 Mortality and Causes of Death Collaborators. 1053
1000 scheme and randomization at the household and com- Global, regional, and national life expectancy, all-cause 1054
1001 munity levels were designed to obtain a representative mortality, and cause-specific mortality for 249 causes of 1055
1002 sample, it is possible that nonresponse bias and com- death, 1980-2015: a systematic analysis for the Global Burden 1056
of Disease Study 2015. Lancet. 2016;388:1459–1544.
1003 munity differences biased our sample. Although all 1057
1004 household locations were mapped across the 2 com- 2. Hill NR, Fatoba ST, Oke JL, et al. Global prevalence of chronic
1058
kidney disease—a systematic review and meta-analysis.
1005 munities, it is also possible that more than 1 household PLoS One. 2016;11:e0158765.
1059
1006 resided at a location (e.g., a compound containing 3. Kearney PM, Whelton M, Reynolds K, et al. Global burden of
1060
1007 several generations of family members), which was not hypertension: analysis of worldwide data. Lancet. 2005;365: 1061
1008 accounted for in our randomization. 217–223. 1062
1009 Our findings reflect a significant burden of NCDs in 4. Whiting DR, Guariguata L, Weil C, Shaw J. IDF Diabetes atlas: 1063
1010 a rural community in Tanzania. Furthermore, aware- global estimates of the prevalence of diabetes for 2011 and 1064
1011 ness of CKD, HTN, and DM was low; the capability to 2030. Diabetes Res Clin Practice. 2011;94:311–321. 1065
1012 identify and diagnose these conditions is largely un- 5. Stanifer JW, Maro V, Egger J, et al. The epidemiology of 1066
1013 available; and adequate treatment of these conditions is chronic kidney disease in northern Tanzania: a population- 1067
based survey. PLoS One. 2015;10:e0124506.
1014 sorely lacking. Although outpatient medical facilities 1068
1015 and health care workers are available in the local 6. Janmohamed MN, Kalluvya SE, Mueller A, et al. Prevalence 1069
of chronic kidney disease in diabetic adult out-patients in
1016 community health centers of Kisarawe, health care Tanzania. BMC Nephrol. 2013;14:1–5.
1070
1017 services are severely resource limited, especially for 1071
7. Pickering TG, Hall JE, Appel LJ, et al. Recommendations for
1018 NCDs. This finding resonates throughout Tanzania, as 1 blood pressure measurement in humans and experimental
1072
1019 recent survey of 24 public and not-for-profit health animals: part 1: blood pressure measurement in humans: a 1073
1020 facilities conducted across Tanzania found primary care statement for professionals from the Subcommittee of 1074
10 Kidney International Reports (2018) -, -–-

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce

 DW Ploth et al.: Chronic Kidney Disease, Diabetes, and Hypertension CLINICAL RESEARCH

1075 Professional and Public Education of the American Heart 25. Yoon SS, Carroll MD, Fryar CD. Hypertension prevalence and 1129
1076 Association Council on High Blood Pressure Research. Cir- control among adults: United States, 2011–2014. NCHS Data 1130
1077 culation. 2005;111:697–716. Brief. 2015;220:1–8.
1131
1078 8. American Diabetes Association. Diagnosis and classification 26. Stanifer JW, Cleland CR, Makuka GJ, et al. Prevalence, risk 1132
of diabetes mellitus. Diabetes Care. 2010;33(suppl 1): factors, and complications of diabetes in the Kilimanjaro re-
1079 S62–S69. gion: a population-based study from Tanzania. PLoS One.
1133
1080 2016;11:e0164428. 1134
9. World Health Organization. Use of Glycated Haemoglobin
1081 (HbA1c) in Diagnosis of Diabetes Mellitus: Abbreviated 27. Mayega RW, Guwatudde D, Makumbi FE, et al. Comparison
1135
1082 Report of a WHO Consultation. Geneva: World Health Orga- of fasting plasma glucose and haemoglobin A1c point-of-care 1136
1083 nization; 2011. tests in screening for diabetes and abnormal glucose regu- 1137
1084 10. Peake M, Whiting M. Measurement of serum creatinine— lation in a rural low income setting. Diabetes Res Clin Pract. 1138
1085 current status and future goals. Clin Biochem Rev. 2006;27: 2014;104:112–120. 1139
1086 173–184. 28. Gallagher EJ, Le Roith D, Bloomgarden Z. Review of hemo- 1140
globin A(1c) in the management of diabetes. J Diabetes.
1087 11. Levey AS, Stevens LA, Schmid CH, et al. A new equation to 1141
estimate glomerular filtration rate. Ann Intern Med. 2009;150: 2009;1:9–17.
1088 1142
604–612. 29. Nitin S. HbA1c and factors other than diabetes mellitus
1089 affecting it. Singapore Med J. 2010;51:616–622.
1143
12. Inker LA, Astor BC, Fox CH, et al. KDOQI US commentary on
1090 1144
the 2012 KDIGO clinical practice guideline for the evaluation 30. Lacy ME, Wellenius GA, Sumner AE, et al. Association of
1091 and management of CKD. Am J Kidney Dis. 2014;63:713–735. sickle cell trait with hemoglobin A1c in African Americans. 1145
1092 13. K/DOQI clinical practice guidelines for chronic kidney disease: JAMA. 2017;317:507–515. 1146
1093 evaluation, classification, and stratification. Am J Kidney Dis. 31. Coresh J, Astor BC, Greene T, et al. Prevalence of chronic 1147
1094 2002;39(2 suppl 1):S1–S266. kidney disease and decreased kidney function in the adult US 1148
1095 14. R: A language and environment for statistical computing population: Third National Health and Nutrition Examination 1149
1096 [computer program]. Vienna, Austria: Foundation for Statisti- Survey. Am J Kidney Dis. 2003;41:1–12. 1150
1097 cal Computing; 2016. Available at: https://www.R-project.org/. 32. Coresh J, Selvin E, Stevens LA, et al. Prevalence of chronic 1151
1098 15. Stata Statistical Software [computer program]. College Sta- kidney disease in the United States. JAMA. 2007;298:
1152
2038–2047.
1099 tion, TX: StataCorp LP; 2015. 1153
33. Barsoum RS. Chronic kidney disease in the developing world.
1100 16. Prevalence of chronic kidney disease and associated risk
N Engl J Med. 2006;354:997–999.
1154
1101 factors—United States, 1999–2004. MMWR Morb Mortal Wkly
1155
Rep. 2007;56:161–165. 34. Kupferman J, Amador JJ, Lynch KE, et al. Characterization of
1102 Mesoamerican nephropathy in a kidney failure hotspot in
1156
17. Inker LA, Shafi T, Okparavero A, et al. Effects of race and
1103 gender on measured GFR: the multi-ethnic study of athero- Nicaragua. Am J Kidney Dis. 2016;68:716–725. 1157
1104 sclerosis. Am J Kidney Dis. 2016;68:743–751. 35. Lunyera J, Mohottige D, Isenburg MV, et al. CKD of uncertain 1158
1105 18. Stanifer JW, Jing B, Tolan S, et al. The epidemiology of etiology: a systematic review. Clin J Am Soc Nephrol. 1159
1106 chronic kidney disease in sub-Saharan Africa: a systematic 2016;11:379–385. 1160
1107 review and meta-analysis. Lancet Glob Health. 2014;2: 36. Shepherd J, Warner MH, Kilpatrick ES. Stability of creatinine 1161
1108 e174–e181. with delayed separation of whole blood and implications for 1162
1109 19. Stanifer JW, Egger JR, Turner EL, et al. Neighborhood clus- eGFR. Ann Clin Biochem. 2007;44:384–387. 1163
1110 tering of non-communicable diseases: results from a 37. Stevens LA, Claybon MA, Schmid CH, et al. Evaluation of the 1164
community-based study in northern Tanzania. BMC Public CKD-EPI equation in multiple races and ethnicities. Kidney Int.
1111 1165
Health. 2016;16:226. 2011;79:555–562.
1112 1166
20. USRDS Report: Incidence, prevalence, patient characteristics, 38. Agoons DD, Balti EV, Kaze FF, et al. Performance of three
1113 & treatment modalities. Am J Kidney Dis. 2014;63:e215–e228. glomerular filtration rate estimation equations in a popula-
1167
1114 tion of sub-Saharan Africans with type 2 diabetes. Diabet 1168
21. Mpondo BCT, Neilson E, Ernest A. Prevalence of chronic
1115 kidney disease in diabetic adult out-patients in Tanzania. BMC Med. 2016;33:1291–1298. 1169
1116 Nephrology. 2016;17:71. 39. Wyatt CM, Schwartz GJ, Owino Ong’or W, et al. Estimating 1170
1117 22. Kavishe B, Biraro S, Baisley K, et al. High prevalence of kidney function in HIV-infected adults in Kenya: comparison 1171
1118 hypertension and of risk factors for non-communicable dis- to a direct measure of glomerular filtration rate by iohexol 1172
1119 eases (NCDs): a population based cross-sectional survey of clearance. PLoS One. 2013;8:e69601.
1173
1120 NCDS and HIV infection in northwestern Tanzania and 40. Samra M, Abcar AC. False estimates of elevated creatinine. 1174
southern Uganda. BMC Med. 2015;13:126. Permanente J. 2012;16:51–52.
1121 1175
1122 23. Hendriks ME, Wit FW, Roos MT, et al. Hypertension in sub- 41. Peck R, Mghamba J, Vanobberghen F, et al. Preparedness of
1176
Saharan Africa: cross-sectional surveys in four rural and Tanzanian health facilities for outpatient primary care of hy-
1123 urban communities. PLoS One. 2012;7:e32638. pertension and diabetes: a cross-sectional survey. Lancet
1177
1124 24. Galson SW, Staton CA, Karia F, et al. Epidemiology of Glob Health. 2014;2:e285–e292. 1178
1125 hypertension in northern Tanzania: a community-based 42. Jha V, Garcia-Garcia G, Iseki K, et al. Chronic kidney disease: Q18 1179
1126 mixed-methods study. BMJ Open. 2017;7(11). Available at: global dimension and perspectives. Lancet. 2013;382: Q17 1180
1127 http://bmjopen.bmj.com/content/7/11/e018829.abstract. 260–272. Q16
1181
1128 1182
Kidney International Reports (2018) -, -–- 11

FLA 5.5.0 DTD
EKIR344_proof
18 May 2018
7:49 pm
ce