Predictors of rapid progression of estimated glomerular filtration rate among persons living with diabetes and/or hypertension in Ghana: Findings from a multicentre study

Abstract In Ghana, the prevalence of chronic kidney disease (CKD) is 28.5% in diabetic hypertensive patients, 26.3% in hypertensives, and 16.1% in those with diabetes only. Trajectories of estimated glomerular filtration rate (eGFR) among patients with hypertension and diabetes are important for monitoring and instituting prompt interventions to prevent the development of CKD, especially in the face of limited access to renal replacement therapy. In this prospective multi‐center study conducted at five hospitals in Ghana, we assessed predictors of rapid eGFR progression among adults with hypertension and/or diabetes. Serum creatinine at baseline and 18 months were taken and eGFR determined using the CKD‐EPI formula. eGFR trajectory was defined as fast when the decline of GFR was ≥ 5 ml/min/1.73 m2 per year. A multivariable logistic regression model was fitted to identify predictors of the fast progression of eGFR. Total 13% of 1261 participants met the criteria for rapid decline in eGFR. The adjusted odds ratio, aOR (95%CI), of four factors adversely associated with fast progression of eGFR were: increasing age 1.20 (1.03–1.14), partial health insurance coverage for medications 1.48 (1.05–2.08), history of smoking 1.91 (1.11–3.27), angiotensin‐receptor blockade use 1.55 (1.06–2.25) while metformin use was protective .56 (.35–.90). Proportion with eGFR <60 ml/min increased from 14% at baseline to 19% at month 18. Effective health insurance policies to improve medication access and avoidance of smoking are interventions that may mitigate the rising burden of CKD in individuals with diabetes mellitus and/or hypertension.


INTRODUCTION
Chronic kidney disease (CKD) is a major public health challenge that contributes significantly to global morbidity and mortality from noncommunicable diseases. The rising burden of CKD is a major drawback to the United Nations target of reducing non-communicable disease by a third by 2030. 1 The major public health threat of CKD is due to increasing risk factors such as hypertension and diabetes. 2 The global prevalence of CKD is between 11% and 13% of the world's population. 3 In Africa, the prevalence of CKD is estimated to be 15.8% among the general population increasing to 34.5% in hypertension and 24.7% in diabetes subpopulations in Africa. 4,5 In Sub-Saharan Africa, the prevalence of CKD is found to be nearly twice higher in high-risk populations with comorbidities than in general populations. 5 The prevalence of CKD in Ghana is 13.3% according to a recent study. 6 Hypertension and diabetes are the leading risk factors for CKD worldwide. 7 In a multicentre study conducted in Ghana, the prevalence of CKD was 28.5% in diabetic hypertensive patients, 26.3% in hypertensives, and 16.1% in those with diabetes only. 2 CKD accentuates cardiovascular mortality approximately eight to tenfold in patients with diabetes and hypertension and is an important contributor to cardiovascular morbidity. 8 In a prospective hospital-based cohort of Ghanaians with hypertension and diabetes mellitus, CKD was found to be independently associated with stroke occurrence in a dose-dependent manner. 9 Renal function trajectory is defined as the change in estimated glomerular filtration rate (eGFR), in ml/min/1.73m2, over time and it is an indicator used in assessing the progression of CKD. 10 The trajectory of eGFR has been studied in different populations with different characteristics. The trajectory of eGFR has been shown to be linear in non-diabetics with a linear decline in renal function. 11 The eGFR of those diagnosed with diabetes declines almost twice as rapidly as those without diabetes. Steeper declines were seen in diabetics who were either Black, had poor glycemic control or had other comorbidities such as hypertension. 12 In another study of people with type 2 diabetes, those with younger age-of-onset or longer duration of type 2 diabetes had a more rapid decline in eGFR compared to those diagnosed in middle age or those with shorter duration of diabetes. 13 This suggest that early and careful monitoring of eGFR changes in those with youngeronset type 2 diabetes is essential as they are at the highest long-term risk for kidney complications.
In Africa, there is a dearth of literature on the trajectory of eGFR among high-risk patients. Such data among people with noncommunicable diseases are particularly important for monitoring and instituting prompt interventions to prevent the development of CKD, especially in the face of limited access to renal replacement therapy and the high cost involved in its management. 12 The lack of reliable data in this area impedes the achievement of the sustainable development goal (SDG) 3.4 target of reducing by third, premature mortality from chronic NCDs through prevention and treatment by 2030. Thus, this study firstly sought to describe patterns of eGFR trajectory over 18 months and secondly to provide information on the factors associated with rapid decline in eGFR among patients with diabetes and hypertension using data from five hospitals in Ghana.

Study design
The Ghana Access and Affordability Program (GAAP) was a prospec-

Eligibility criteria
The inclusion criteria were adults (i.e., aged 18 years and above) with a known diagnosis of hypertension and or type 2 diabetes mellitus.
Hypertension was defined as a persistently elevated systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg in patients who were on antihypertensive drugs at the enrollment visit.
Diabetes mellitus was defined as a fasting blood glucose ≥7 mmol/L on two or more occasions, random blood glucose ≥11.1 mmol/L with symptoms, hemoglobin A1C (HbA1C) of > 6.5%, and on medications for the treatment of type 2 diabetes mellitus. Exclusion criteria included individuals with hypertensive urgencies or emergencies, hyperglycaemic or hypoglycaemic emergencies, and type 1 diabetes mellitus.

Evaluation of study participants
All eligible respondents provided informed consent before enroll-

Laboratory measurements
An International Organization for Standardization (ISO)-certified and quality-assured laboratory was contracted to run all biochemical panels which included serum creatinine, and hemoglobin A1C for participants with diabetes. Samples were transported to the laboratory by trained phlebotomists on the same day of collection often within 4 h or where not feasible (Atua Government Hospital and Kings Medical center), samples were stored in a freezer before being transported to the laboratory the next day. The eGFR was determined from serum creatinine using the CKD-EPI formula. 14 CKD was defined as either eGFR < 60 ml/min/1.73 m 2 with or without trace proteinuria and above or eGFR ≥60 ml/min with proteinuria according to the Kidney Disease Improving Global Outcome. 15 Serum creatinine at baseline and the end of the study at 18 months were taken.

Primary outcome
Rapid decline of eGFR defined as more than 5 ml/min/1.73m 2 per year. 16

Secondary outcome
The frequency or proportion of study participants according to categories of eGFR trajectory into five patterns, namely, rapid decline, slow decline, stable, slow improvement, rapid improvement. This was based on the following definitions:

Baseline characteristics of study participants
The study included 1261 participants from the 5 study sites.

Mean eGFR changes from baseline to month 18
There was an overall decline in the mean eGFR from 77. 22

Trajectory of eGFR patterns over 18 months
Of the study participants, 13% had a rapid decline in eGFR, 10% had a steady decline, 30% had stable eGFR or no change in eGFR, 16% had steady improvement, and 31% had rapid improvement in the eGFR over 18 months. Furthermore, while at baseline 42% of participants had eGFR > = 89 ml/min, at month 18 the proportion at this eGFR dropped to 32%. However, those with eGFR 60-89 ml/min at baseline increased from 44% to 48%, eGFR 30-59 ml/min increased from 9% to 14%, eGFR 15-29 ml/min increased from 1% to 4% with no changes in eGFR < 15 ml/min (0% vs. 0%). (Table 1) Overall, the proportion with eGFR < 60 ml/min increased from 14% at baseline to 19% at month 18.

Comparison of clinical and demographic characteristics of study participants by progression of eGFR
There were 164 (13.0%) with rapid progression of eGFR. The mean age in years of participants with rapid eGFR progression was higher compared to those with normal progression [60.3 ± 10.5 vs. 57.1 ± 11.8; (p = .001)] Table 2. For participants whose medications were fully paid for by the National Health Insurance Scheme, a significantly lower proportion had fast progression of eGFR compared with those whose  Table S1.
Furthermore, among those with both hypertensin and diabetes, three factors independently associated with rapid decline in eGFR were partial insurance coverage of medications, smoking history and duration of diabetes diagnosis ( Table 5). None of the factors in those with hypertensives only was independently associated with rapid progression (Table S2).

DISCUSSION
We found in this prospective multi-centre Ghanaian study that 14% Admittedly these cited studies involved a heterogeneous study sample and eGFR decline rates were different across the studies. We choose a less conservative cut-off of ≥5 ml/min/1.73 m 2 per year. Even so, we observed a significant burden of rapid declines in kidney function among Ghanaians living with hypertension and or diabetes compared with previously cited studies. [17][18][19] In the present study, we found increasing age, partial health insurance coverage for medications, history of smoking, and prescription of ARB were associated with a rapid decline in eGFR while patients on Metformin had a slow progression. Each 10-year rise in age was associated with 20% increased odds of faster progression of eGFR.
Age is an important non-modifiable risk factor that impacts the trajectory of eGFR. Both morphological and functional changes occur in the senescent kidney. 20 Global glomerulosclerosis increases steadily with increasing age in the absence of any existing co-morbidity such as hypertension or diabetes and there is a gradual decline in the overall number of nephrons as one age. 21 Overall, the age-related decline in eGFR usually begins after the third decade of life and most often progresses at a rate of 8 ml/min per 1.73 m 2 per decade 22-24 which usually accelerates after the age of 50-60 years. 25 The progress of this process may be influenced by superimposed diseases such as diabetes and hypertension. 26 Diabetes, hypertension and aging have been found as independent key predictors of CKD. 27 As noted in a previous study, we observed a higher rate of progression of eGFR among individuals with partial insurance coverage for Abbreviations: ACE, angiotensin converting enzyme; ARB, angiotensin receptor blocker, BMI, body mass index; DD4 = Dipeptidyl peptidase-4; DM, diabetes mellitus; NHIS, National Health Insurance Scheme. p-value < .005 is significant. Table 2 indicates the clinical and demographic characteristics of all patients in the study. Advanced Age, drug paid by NHIS, disease class, history of smoking, ARB use, number of antidiabetic medications, metformin use are the significant characteristics that affect the patient's eGFR.
antihypertensive and or antidiabetic medications. 28 Access to essential medicines to manage non-communicable diseases is severely limited in public sector settings in Low-and Middle-Income Countries and this adversely impact disease outcomes. 29,30 In this cohort, we have shown that the absence of National Health Insurance Scheme (NHIS) was adversely associated with poor glycemic control. 31 Also uncontrolled hypertension was associated with poor adherence to antihypertensive therapy due to reported difficulty in obtaining medications in this cohort. 32 We found that a history of cigarette smoking was associated with a higher adjusted odds of rapid eGFR progression by 91% compared with those who did not smoke. A meta-analysis showed cigarette smoking, that is, ever smokers, current smokers, and former smokers as an independent risk factor for incident CKD. 33 Although smoking is dose-dependently related to incident kidney failure, prolonged cessation of smoking significantly reduced this risk. 34,35 Also salt intake was found to be significantly associated with rapid progression of eGFR in patients with diabetes. Salt added to food increases urinary albumin excretion which significantly increases risk of cardiovascular events and progression of diabetic nephropathy. [36][37][38] An intriguing observation was that of a faster progression of eGFR among participants on ARBs. A possible explanation could be indication bias where ARBs were initiated by clinicians among patients who already have proteinuria to avert progressive renal function decline. 39 On the other hand, in line with evidence from previous studies, metformin had a protective effect on the decline in eGFR. 39 Table 3, the determinants of rapid progression in eGFR reported for all study participants as adjusted odds ratio (aOR) with 95%CI include increasing age per 10 years, partial NHIS drugs payment, ever smoked before and ARB use. Patients on metformin were less likely to have fast progression of eGFR.  Table 5 shows that none of the factors in those with hypertensives only was independently associated with rapid progression.

Implications
Our study has implications for individuals living with hypertension and diabetes, health care providers, and policymakers/directors. At an individual patient level, adherence to therapeutic lifestyle interventions such as smoking cessation and medications prescribed by clinicians would improve blood pressure and glycemic control and avert decline in renal function. Given the observed association between increasing age and risk of rapid progression of eGFR, the ageing population liv-ing with diabetes or hypertension and other high CV risk groups [41][42][43][44] should be educated on the need for regular screening of renal function.
For clinicians, more frequent assessment of eGFR among hypertensive and diabetics perhaps yearly may help with the identification of individuals at risk of rapid progression of eGFR, for referral to a nephrologist or a specialist. While difficult to explain the association of the use of ARBs and the rapid progression of CKD, those on ARBs may require more frequent monitoring. At the policy level, an expanded and affordable health insurance policy to cover all medications for patients with hypertension and or diabetes will undoubtedly improve access and affordability to these lifesaving medications to prevent end-organ complications such as end-stage kidney disease.

Limitations
This prospective cohort study had some important limitations. Estimated glomerular filtration rates were assessed at two-time points, baseline and 18 months later and hence we could not evaluate fluctuations in eGFR in between these two-time points. Also, there could be selection bias as only participants who had both baseline renal function and end of study renal function were included in this study. Despite these limitations, this present study fills an important gap in the literature by identifying predictors of rapid decline in renal function among individuals with diabetes and or hypertension in resource-limited settings in Ghana.

CONCLUSION
The trajectory of eGFR is affected by both modifiable and non-

ACKNOWLEDGMENTS
Funding for this study was provided by MSD, Novartis, Pfizer, Sanofi, and the Bill and Melinda Gates Foundation (collectively, the Funders) through the New Venture Fund (NVF). The funders had no role in study design, data collection, data analysis, or study report writing.

CONFLICT OF INTEREST
Authors declare no conflict of interest.