Analytical study with follow-up of two dynamic cohorts of CKD patients to compare the development of clinical indicators, laboratory values, and indicators of renal impairment in patients exposed to a RPP versus patients on conventional therapy (CT).

The population studied was the census of users of two Colombian HPEs. Retrospective information obtained from the electronic history records (HER) of both insurance companies was collected from the year the RPP started until the moment the study started. Then the information was obtained prospectively until completion of a five year follow-up. The inclusion criteria were: being affiliated to the HPE during follow-up and meeting the KDOQI clinical practice guidelines for CKD.16,17 Children under fifteen years of age and those registered in both institutions at the same time or in another during the period of study were excluded. The beginning of the follow-up was set at the moment of diagnosis of CKD or at the moment of inclusion in the RPP. Ethical support from the IRB/IEC of the institutions involved was granted.

The RPP is a secondary prevention strategy that includes a protocol of educational and clinical appointments, with clinical and periodic lab exams focused on the active search for patients at risk. The RPP has a first level of healthcare for the follow-up of patients in the early stages of CKD with general practitioners (GP) and nurses and two annual appointments with the Internal Medicine and Nutrition units. The RPP has a second level of healthcare from CKD stage 3 or more of with appointments with the Internal Medicine, Nephrology, and Nutrition units every three or two months based on the stage of CKD. The lab exams are conducted one month prior to the appointment following the RPP technical guide.

The other group is on CT which provides first level healthcare both for patients at risk and for those with a diagnosis of CKD; referral to a specialist occurs based on the medical criterion from the treating GP and as requested by the patients. No guidelines or protocols of care are applied, and there is no active search for patients at risk.

Variables based on exposure (based on being exposed to the RPP or CT). The variables considered as criteria for the diagnosis of CKD were: creatinine clearance <60ml/min (estimated using the Cockroft–Gault formula), alterations in renal ultrasound, proteinuria >150ml/day, abnormal urinary sediment and, in diabetic patients microalbuminuria >30mg.16,17

Dependent variables: time from diagnosis to first stage progression, and time from diagnosis to need for RRT.

Independent variables: the clinical and lab characteristics.

Information was gathered from databases provided by every HPE and data obtained from the active and progressive search from clinical charts of patients during the period of study.

An analysis of the population census of both insurance companies meeting the diagnostic criteria based on the KDOQI clinical practice guidelines was conducted.

Quantitative variables: the clinical markers such as of systolic blood pressure, diastolic blood pressure, fasting glycemia, glycosylated haemoglobin (HbA1c), total cholesterol, triglycerides, low and high density lipoproteins (LDL, HDL) and creatinine were dichotomically recodified as controlled/non-controlled variables based on what it is recognized by the scientific literature for each clinical parameter in CKD patients. The reference values to define a variable as controlled were: systolic blood pressure <130mmHg, diastolic blood pressure <80mmHg, glycemia <140mg/dL, HbA1c <7 per cent, total cholesterol <200mg/dL, triglycerides <150mg/dL, LDL <100mg/dL, HDL >40mg/dL and creatinine <1.2mg/dL.16,17 The glomerular filtration rate (GFR) was considered a progression variable based on each CKD stage. In some of the analyses it was preferred to group variables such as age, marital status, stage of the disease, total number of appointments, and clinical markers.

Population was characterized for the diagnosis of CKD through frequency distributions and proportions for qualitative variable; for the quantitative variables, we used descriptive statistics and dispersion statistics; in addition, the type of distribution was assessed using the Kolmogorov–Smirnov or Shapiro–Wilk normality tests. In order to establish the differences between the RPP and CT, the Student's t test or the Mann–Whitney U test were used based on the distribution of the variables.

The rate of the different outcomes in RPP and CT was estimated. To this end, in the numerator was the number of new cases of first progression and requiring RRT, and in the denominator, the years of exposure from the moment of diagnosis of CKD. Also the ratio of the incidence rate with its corresponding confidence intervals (CI) at 95 per cent was estimated.

The Kaplan–Meier survival curves were used to assess stage progression and need for RRT per insurance company. Both models were adjusted for the different covariables and also for two subgroups based on stage: 1 and 2 as early entry, and 3 and 4 as late entry. Stage of disease, age, and gender were all analyzed comparatively (log rank test) in an effort to compare the progression in each stage. Cox proportional hazards analysis was used to analyze the influence of the exposure to RPP on each dependent variable adjusted for the other variables of interest. To compare progression, the late stages 3 and 4 were considered reference points since progression is lees, and stage 5 was not taken into account because it does not progress any further. On the contrary, to compare the need for RRT, the early stages 1 and 2 were considered reference points since late stages represented a too high risk of RRT. Covariables were included in the models based on clinical and statistical criteria (Hosmer–Lemeshow: p<0.25). Compliance with the proportional hazards assumption was tested graphically using Scaledsch and Schoenfeld scaled residuals in Stata Version 11.

There were 9887 clinical histories consulted; 56.5% corresponded to patients within the RPP and 43.5% to patients on CT. Overall, 5663 patients meet the inclusion criteria, 74.2% had been exposed to the RPP and 25.8% had been on CT. During follow-up there were 509 withdrawals (8.9%), 1010 losses to follow-up (17.8%), and 312 deaths (5%). The age of patients within the RPP was significantly lower than the age of patients on CT, with an average age of 66 (16–100) and 70 (20–97) years old, respectively. There were more male patients in the group on CT (64.5%; p<0.001).

The most frequent co morbidities in both groups were the following: high blood pressure (HBP) (94.67% for the RPP and 85.01%for CT; p<0.001); dyslipidemia (60.42% for the RPP and 63.8% for CT; p=0.023) and diabetes mellitus (DM) (36.2% for the RPP and 32.65% for CT; p=0.014). More than 30% of patients from the RPP were stage 1 and 2 patients while only 16.7% of patients on CT were stage 1 and 2 patients.

As compared to patients on CT, patients on RPP took longer to progress from one stage of the disease to the next, they expend more time until their first hospitalization, and also it took more time to require RTT. All these differences were clinically relevant and statistically significant (Table 1).

In patients on RTT, a 24% progressed to next stage and in CT only 17% showed progression of disease stage (p<0.001). On the other hand, the percent of patients requiring RTT after follow-up was higher in patients on CT (5.1%) than in RPP (4.8 per cent) but p=0.802.

Patients on the RPP showed more progression in the early stages of the disease. However, in later stages of CKD, both cohorts showed a tendency to stay in a given stage; but, patients within the RPP have a better control of renal function since a high percentage of these patients had a regression of the stage (Fig. 1).

Fig. 1.

Stage variation between diagnosis and end-stage in patients with CKD.

(0.25MB).

The impairment of GFR was less in RPP than CT at the moment of diagnosis, and with the first and last changes of stage (Table 2). Among patients that progressed during follow-up, we found that patients on RPP showed more favourable GFR at the moment of diagnosis and during the first and last changes of stage than patients on CT. Such differences are clinically relevant and statistically significant (Mann–Whitney U test p<0.001 for the three instances).

In the group of patients who needed RRT during follow-up, we saw that patients within the RPP showed better glomerular filtration rates both at the moment of diagnosis and during the first and last changes of stage than patients on CT, though only the first of such differences was statistically significant (Mann–Whitney U test with p=0.022; 0.191; 0.051, respectively).

At the end of the study the incidence rate (IR) of stage progression and need for RRT was recorded in both groups. RRT was needed in 204 patients on RPP, and in 74 patients on CT. Thus the IR for progression was greater for CT patients (0.050; 95% CI: 0.040–0.064) than RPP patients (0.034; 95% CI: 0.030–0.039). The IR ratio for such outcome was 1.48 (95% CI: 1.21–1.90). The IR for stage progression was not significantly different between the groups.

Kaplan–Meier curves adjusted for subgroups and other covariables showed that in both models the patients on RPP had less stage progression and less need for RTT than patients on CT (Figs. 2 and 3).

Fig. 2.

Survival curve for stage progression in RPP vs. CT.

(0.07MB).

Fig. 3.

Survival curve for need for RTT in RPP vs. CT.

(0.06MB).

Analysis of stage progression: hazards model adjusted for stage subgroups and different covariables showed that patients on RPP were significantly protected from stage progression (HR: 0.855; 95% CI 0.740–0.988). It was confirmed the influence on stage progression of age (HR: 1.623; 95% CI: 1.429–1.843), HBP (HR: 1.55; 95% CI: 1.209–1.988) and DM (HR: 1.216; 95% CI: 1.085–1.362). Also, patients in early stages of the disease had higher hazard ratios of stage progression (HR: 4.039; 95% CI: 3.554–4.590) while dyslipidemia turned out to be protective (Table 3).

The hazard ratio for RRT was lower in the RPP (HR: 0.797; 95% CI: 0.606–1.049) although it was not statistically significant. However patients in the late stages of the disease showed a statistically significant increased hazard ratio for RRT (HR: 4.909; 95% CI: 3.10–7.776). Comorbidities like DM increased the hazard ratios for RRT (HR: 2.080; 95% CI: 1.637–2.643). Age and gender predisposed to a higher HR for RRT but they were not statistically significant. In addition dyslipidemia reduced the HR for RRT (Table 4Table 4).