Non-linear progression of chronic kidney disease and associated factors in hypertensive patients: a 4-year cohort study

Recent studies suggest that the progression of chronic kidney disease (CKD) is not linear, but we do not have clear evidence on this issue, especially in hypertensive patients. We sought to evaluate the progression of CKD and associated factors over four years in a cohort of hypertensive patients. Methods We conducted a prospective cohort study during the years 2012 and 2016, with hypertensive patients diagnosed with CKD (n = 113). The progression of CKD was assessed through the evolution of the glomerular ltration rate (GFR) and the change in the stage of CKD between 2012 and 2016. Sociodemographic, economic, lifestyle, clinical, anthropometric, and biochemical variables were evaluated. The strength of the association between CKD progression and explanatory variables was assessed by odds ratio (OR) and their respective 95% condence intervals using univariate and multivariate logistic regression.


Introduction
Traditionally, chronic kidney disease (CKD) is characterized by a progressive loss of renal function, often manifested by a decline in glomerular ltration rate (GFR), over a long period of time [1]. However, recent studies [2][3][4] suggest that the progression of CKD is not linear. Many patients suffer from the nonprogressive or slowly-progressive form of the disease and will never develop terminal kidney disease. Conversely, some patients evolve rapidly to terminal kidney failure [4].
To develop effective strategies to delay or prevent the progression of CKD, it is important to evaluate the changes that occur in the GFR over time, in order to know more about the disease history. Previous studies on the prognosis of CKD have not provided clear evidence on this issue. Epidemiological studies on CKD were not longitudinal [7,8], and the longitudinal data did not provide estimates of change in GFR [9][10][11][12][13]. In addition, some studies have not veri ed if the reduced GFR persisted for three months or more, as required by the de nition of CKD [8,10], thus generating bias by the inclusion of patients with reversible acute renal insu ciency. Other studies [3,4] did not show the evolution of GFR through repeated pre-scheduled measures for the same time interval during the monitoring period, which would be the ideal design according to Eriksen and Ingebretsen [14]. Finally, recent studies have not investigated the factors associated with the progression of CKD [2][3][4]15].
Following what was proposed as the ideal design, the objective of this longitudinal observational study was to evaluate the progression of CKD and associated factors over four years in a cohort of hypertensive patients.

Study design and subjects
We conducted a prospective cohort study during the years 2012 and 2016, with hypertensive patients accompanied by primary health care teams from the municipality of Porto Firme, Minas Gerais, Brazil.
The inclusion criteria in the study were individuals aged 18 years or older with AH and who agreed to participate in the study after due clari cation. The exclusion criteria were individuals with severe clinical conditions who needed specialized care, pregnant women, individuals with a history of alcohol and/or drug abuse, and individuals with a diagnosis of established CKD.
At baseline, there were 697 hypertensive patients registered in the primary care information system. The sample was de ned considering the reference population with AH in 2012 of the municipality (n=697), a prevalence of 50% of the phenomenon studied, 5% sample margin of error, 10% of refusals and/or losses, 20% to control confounders, and 95% con dence level. The sample calculation resulted in a minimum sample of 248 individuals. The nal study sample consisted of 293 individuals. The sample calculation was performed using the Statcalc program of Epi-Info® version 7.2.
A previous study [16] screened 293 individuals for CKD through the GFR estimated by the formula Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) and by analysis of proteinuria and albuminuria in urine of 24 hours. Of these, 113 individuals were diagnosed with CKD. CKD patients were individuals with GFR < 60 mL/min/1.73m², with con rmation three months after the initial diagnosis.
In the present study, patients diagnosed with CKD (n = 113) were followed for four years. At the end of the follow-up, there were 96 patients with CKD, which corresponds to 85% of those initially evaluated. The loss of 17 individuals was due to death (n=14), change of town (n=3), and declined participation (n=4).
During the follow-up, six measurements of GFR and albuminuria were performed. The progression of CKD was assessed through the evolution of the GFR and the change in the stage of CKD between 2012 and 2016.

Data collection
Data collection occurred through individual interviews and anthropometric and biochemical evaluations. As an instrument of data collection, a semi-structured interview guide was used, addressing sociodemographic and economic (gender, age, civil status, education, and family income), lifestyle (tobacco use, alcohol intake, and physical activity), and clinical (DM and time with AH) variables.
Physical activity was assessed using the short version of the World Health Organization (WHO) International Physical Activity.
In relation to anthropometric measurements, weight, height, and waist circumference (WC) were assessed. The weight was obtained by means of electronic balance, with a capacity of 150 kg and division of 50 grams; the height was measured using a portable anthropometer, composed by a metallic platform for positioning the individuals and a removable wooden column containing a millimeter tape and cursor for reading, according to the techniques proposed by Jelliffe [17]. The body mass index (BMI) was calculated by the ratio between weight and squared height, and classi ed according to the criteria of the WHO [18] for adults and Lipschitz [19] for the elderly. The WC measurement was performed using an inextensible tape and measured in centimeters, at the midpoint between the iliac crest and the external face of the last rib. The values were classi ed in relation to the risk for cardiovascular diseases and metabolic complications according to the cutoff points proposed by the WHO [18].
Regarding the biochemical variables, serum creatinine and albuminuria were evaluated to analyze renal function. Albuminuria was considered abnormal when the value was ≥ 30 mg/24h, as described by Kidney Disease Improving Global Outcomes (KDIGO) [1]. The GFR was estimated from the CKD-EPI formula, currently recommended by the KDIGO [1] and the Brazilian Ministry of Health [20]. CKD stages were classi ed as follows On the scheduled day, participants attended the laboratory accredited for the blood and urine collection. The participants were instructed to maintain the usual diet during the day and fast for 12 hours before collection. The analysis of the biological material was performed in a private laboratory, using commercial kits.

Statistical analysis
The normality of the data was tested using the Kolmogorov-Smirnov test. Continuous variables were described as mean and standard deviation, and categorical variables were expressed as frequencies.
Differences in baseline characteristics were tested for statistical signi cance with a t test for continuous data and chi-square test for categorical variables. The strength of the association between CKD progression and explanatory variables was assessed by odds ratio (OR) and their respective 95% con dence intervals using univariate and multivariate logistic regression. For data analysis, the software SPSS Statistics for Windows (version 20.0) was used, and the statistical signi cance was set to p ≤ 0.05.

Results
In relation to the sociodemographic and economic characteristics of the studied population, the majority was female, elderly, living with a partner, and with low education and low income. Regarding lifestyle, 67.7% had never smoked, 94.8% did not use alcohol, and 58.3% were physically active. Most participants did not have DM and discovered AH less than 10 years ago. Regarding nutritional status, 61.5% were overweight and 85.4% had cardiovascular risk. The biochemical parameters are described in Table 1. Regarding progression, 78.1% (n=75) of the CKD patients did not progress over four years. There was a statistically signi cant difference between the progression and non-progression groups for DM (p=0.042) and urea (p=0.029). Table 1 In Table 2, when assessing the CKD trajectory (2012-2016) through the evolution of GFR, there was a mean reduction of 1.3 mL/min/1.73m² in four years, which corresponds to 0.32 mL/min/1.73m² per year. In the group that progressed, there was a reduction of 13 mL/min/1.73m² in four years or 3.25 mL/min/1.73m² per year in the GFR, while in the group that did not progress, there was an increase of 2 mL/min/1.73m² in four years or 0.5 mL/min/1.73m² per year, indicating a non-linear trajectory of the GFR over time. It is noteworthy that 12.5% (n=12) of the participants showed rapid progression of CKD over a year (reduction in GFR greater than 3 mL/min/1.73m² per year). Table 2 Graph 1 illustrates the non-linear trajectory of the GFR during the years 2012 and 2016 in the progression and non-progression groups. In 2016, it is possible to observe an increase in the GFR for the nonprogression group.

Graph 1
In the multivariate analyses, presented in Table 3, age ≥ 75 years (p=0.047), DM (p=0.042), and urea (p=0.050) were independently associated with CKD progression. Table 3 Discussion The ndings of the present study showed a non-linear progression of CKD over the four years. During follow-up, there was a reduction of 1.3 mL/min/1.73m² in GFR. In the progression group, there was an average reduction of 13 mL/min/1.73m², which characterizes rapid progression. On the other hand, in the non-progression group, there was an increase of 2 mL/min/1.73m² in GFR, contrary to what is traditionally expected, which is a GFR reduction (slow or fast) or maintenance.  [21], similar to the study by Inaguma et al. [22] that showed an annual average decrease of 0.36 mL/min/1.73 m 2 in GFR. Another study with a median follow-up of 31 months (2.5 years) with patients with estimated GFR less than 30 mL/min/1.73m² showed a decline in GFR of 2.65 mL/min/1.73m² per year [23]. Therefore, different situations are observed in relation to GFR over time, and it is not possible to establish a pattern. In the present study, we can establish two situations: a group that progressed rapidly (3.25 mL/min/1.73m² per year) and another group that maintained a similar GFR and/or increased the GFR in relation to the baseline (non-progression).
Thus, an important nding of this study was the non-linear variation in the values of GFR throughout time, with increased or decreased GFR. The traditional paradigm of GFR progression among patients with CKD shows a steady and almost linear decline over time [3,14]. In contrast to the traditional paradigm, many patients with CKD have a fast and/or slow progression of the disease or an extended period of nonprogression. In the Modi cation of Diet in Renal Disease study, approximately 19% of patients with GFR between 25 and 55 mL/min/1.73m² experienced improvement or stabilization of renal function during the study period of two years [24], which corroborates other studies of great relevance [3,4,14,15,25].
In addition to the individual biological variation, the analytical variation inherent to the estimation of the GFR from serum creatinine measurements should be considered. Some problems related to determining the serum creatinine are inconstant production of serum creatinine, increasing with intake of meat, creatine, or with excessive muscular effort; the Jaffé creatinine analysis method suffers interference, in vitro, positively from cephalosporins and ketone bodies and negatively from bilirubin; enzymatic methods suffer interference, in vitro, from n-acetylcysteine and dipyrone; and creatinine values vary with the presence of simple infections, dehydration, and use of nephrotoxic drugs [26]. Thus, one of the challenges in the use of routine analyses of creatinine to estimate changes in GFR over a long period is to ensure the stability of the test. Although the strict quality control routines protect against major uctuations, variation in the long term cannot be completely excluded [14].
This study also associated the progression of CKD with sociodemographic, economic, lifestyle, clinical, anthropometric, and biochemical variables. Knowing the associated factors may contribute to the correct identi cation of CKD and implementation of actions to slow its progression [27]. In the multivariate analyses, age ≥ 75 years, DM, and urea were independently associated with CKD progression.
In the present study, individuals aged ≥ 75 years were eight times more likely to have CKD progression than individuals aged ≤ 68 years. Age is recognized as an independent risk factor for CKD, and the ndings of association of this disease with aging are consistent with previous studies [28,29]. However, in an analysis of a large cohort, older age was associated with slower loss of GFR, although this seemed to be true only at GFR levels < 45 ml/min per 1.73 m² [30]. Conway et al. [31] described similar results in an elderly population with stage 4 CKD. Thus, it appears that there is some other pathological process, likely vascular in etiology, that plays a role in pathogenesis of GFR decline in the elderly [32]. Considering that all individuals in this study have AH, this factor may also be contributing to the progression of CKD. Data from the Boston Longitudinal Study of Ageing suggest that the decline in GFR with increasing age is largely attributable to AH or the presence of comorbidities, such as heart failure and other cardiovascular diseases [33].
Another important factor is DM, which is the leading cause of CKD in the developed world, and people with diabetes and CKD have a greatly increased risk of all-cause mortality, cardiovascular mortality, and kidney failure [34]. In individuals with diabetes, both GFR and albuminuria are important predictors of kidney outcomes. In a recent meta-analysis, compared to the nondiabetic participants, those with diabetes showed a borderline increased hazard for the progression from late-stage CKD to end-stage renal disease (HR 1.16, 95% CI 0.98-1.38) [35]. On the other hand, the study by Levin et al. [23] had the largest study population and enrolled stage 4 to 5 CKD patients, for whom diabetes was reported as a nonsigni cant factor (HR 0.82, 95% CI 0.56-1.20; P = 0.30). Previous studies have demonstrated large variation in GFR progression in persons with diabetes [36]. In the present study, individuals with diabetes were 8.74 times more likely to have CKD progression than individuals without diabetes.
In relation to urea, individuals with higher urea levels were 8% more likely to have CKD progression. Its elevation indicates inability of the renal system to purify the blood of nitrogenous products. Fehrman-Ekholm and Skeppholm [37] found a signi cantly positive correlation between age and urea (p = 0.0019); that is, urea increased with age, which probably re ects the decrease in GFR, which corroborates the data from the present study. Levey [38] highlights that urea is the rst used endogenous marker, but it is not completely reliable, as its levels are more vulnerable to changes for reasons unrelated to GFR.
Finally, in individuals who are progressing, the subsequent risk of morbidity and mortality increases exponentially, as well as the costs associated to health. A reduced GFR also associates with a wide range of complications and reduced quality of life. Therefore, it is important to clarify which factors are associated with the CKD progression and are potentially modi able, in order to intervene early and improve the associated adverse results [25].
There are limitations to this study. There are medications that may be more commonly taken by hypertensive patients that may have an effect on renal function. We were unable to include these in our analysis. Another limitation of the study was the di culty maintaining consistent calibration of the serum creatinine test over time, and the results are highly sensitive to progression adrift in the assay of creatinine. Despite this, the ndings of this study have implications for future researches on CKD. Identifying non-linearity in the progression of the disease suggests a new approach to other studies that investigate the association of risk factors that vary in time with changes in renal function. Thus, a more accurate understanding of the CKD trajectory can help to guide clinical decision making, develop actions of prevention of disease progression, and seek to improve patients' quality of life.

Conclusion
In conclusion, the ndings of the present study showed a non-linear progression of CKD over the four years, contrary to what is traditionally expected. We can establish two situations: a group that progressed rapidly and another group that maintained a similar GFR and/or increased the GFR in relation to the baseline (non-progression). Age, DM and urea were independently associated with CKD progression.  Informed consent was obtained from all individual participants included in the study. As per Resolution 466/2012 of the National Health Council, which regulates research involving human subjects, the participants signed a free and informed statement of consent, which ensured the con dentiality of the data and the anonymity of the participants.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This project received support from the Foundation for Research Support in the State of Minas Gerais, Brazil (FAPEMIG -process no. CDS-APQ-03594-12).
Authors' contributions LSS participated in the design of the study, the collection of the data, the statistical analysis and interpretation of data, the redaction of the article. TRM participated in the design of the study, the statistical analysis and interpretation of data, the redaction of the article. RGS and RMMC participated in the design of the study, the redaction of the article. All authors read and approved the nal manuscript.