REVIEW FREQUENCY AND OUTCOMES IN CKD STAGE G5: A POPULATION-BASED COHORT IN ENGLAND

KDIGO guidelines recommend eGFR be monitored ≥3 times per year for CKD G4 and ≥4 for G5. However, little is known about how monitoring is conducted in practice. This study aimed to investigate nephrology specialist review and eGFR monitoring frequency of people before and after a diagnosis of CKD G5 and to examine associations with clinical outcomes.

In this study, we used two linked nationwide databases in England, Clinical Practice Research Datalink covering primary care and Hospital Episode Statistics covering secondary care, to capture community and specialist care events of patients with CKD G5. The day of reaching CKD G5 (index date) was defined by two eGFR <15ml/min/1.73 at least 90 days apart, or an eGFR <15ml/min/1.73 with no future ≥15ml/min/1.73 between 2013 and 2018. The 720-day period before and after the index date was divided into eight quarters (each comprising 90 days). For each quarter, we assessed whether patients had i) an eGFR test in primary care, and ii) review by a nephrology specialist. The proportion was defined as the number of quarters with ≥1 monitoring event divided by the total observed quarters. Patients with a proportion ≥0.75 (≥3 monitoring events per year, consistent with guideline recommendations) were classified as the high-monitoring group, while all others were classified as the low-monitoring group. Follow-up continued until kidney replacement therapy (KRT) initiation, death, end of follow-up in the database, or 720 days, whichever occurred first. Logistic regression models were used to assess factors for low monitoring and to examine the associations between monitoring frequency and KRT or death without KRT, adjusted for age, sex, ethnicity, comorbidities, electronic frailty index, eGFR, and deprivation.

5,182 patients with CKD G5 were included in the analysis. Among the high-monitoring (n = 3,125, 42.0% female) and the low-monitoring group (n = 2,057, 47.3% female), the median observation duration was 1092 days (377 days after index and 720 days before). During the observation period, 912 patients died, and 2,707 initiated KRT. Figure 1 illustrates monitoring patterns across the entire follow-up. In the high-monitoring group, specialist reviews increased during the period preceding CKD G5 (−8 to −1), and the proportion of monitored quarters remained consistently high after the index date (+1 to +8). In contrast, the low-monitoring group showed persistently low monitoring rates before CKD G5, with only a limited increase after diagnosis and a tendency to decline over time. Figure 2 shows factors associated with low monitoring: female sex, age ≥80 years, higher deprivation, dementia, and cerebrovascular disease. Figure 3 demonstrates the non-linear association between monitoring frequency and outcomes, showing that higher monitoring frequency was associated with lower odds ratios of death and higher odds ratios of KRT initiation.

We have described how monitoring for people with CKD G5 does not adhere to guideline recommendations. Factors associated with less monitoring include female sex, older age, and higher levels of deprivation, as well as the presence of dementia and cerebrovascular disease. Less monitoring frequency was associated with lower incidence of KRT initiation, and higher incidence of death without KRT. Further research is needed to provide optimal monitoring in patients with CKD.