In this primary care cohort of incident CKD, the risk of CVE and mortality differed according to previous HTN or/and T2D diagnosis. After adjusting for multiple factors associated with HTN, T2D and CKD, the risk of death was higher in the T2D group than in the HTN-CKD group, and it was in-between for HTN/T2D-CKD. For CVE, the risk was similar in HTN and T2D-CKD, and increased when both were present.
The prevalence of HTN in T2D patients in the cohort was high (89.0%), and therefore the percentage of patients in the T2D-CKD group was low (3.87%). This is within the range of similar cohorts in primary care; in patients with T2D and CKD, a prevalence of HTN of 85.4% 15 and 88.6% has been described 16.
The characteristics of individuals differed according to CKD groups. Those with isolated T2D were the youngest, with higher prevalence of male sex and smokers, and lower prevalence of obesity compared to the HTN/T2D-CKD group. Both groups presented a high prevalence of comorbidities, HF and anaemia. On the contrary, individuals in the HTN-CKD were the oldest, more frequently female and had the lower prevalence of smokers and comorbidity.
Although the risk of CKD increases notably with a T2D duration of 10 years or more 15, renal disease evidence was found in approximately three quarters of this cohort < 10 years from HTN or T2D diagnosis, and in more than a third within 5 years. The time span was shorter when T2D presented without evidence of HTN. This could suggest a more aggressive clinical course or reflect a distinctive patient profile for isolated T2D which presents in younger patients. Moreover, individuals who died or developed a CVE had a shorter time of T2D and HTN evolution. Therefore, a shorter time to CKD appearance could identify individuals at higher risk of unfavourable outcomes, and further analysis should be performed.
The higher mortality risk associated with CKD is well known 1,3 and has been confirmed in this population 17,18.
Mortality risk in CKD individuals with/without T2D has been specifically compared. In a previous study, both moderate CKD (eGFR 30–50 mL/min), and T2D similarly increased the risk of death (HR 1.40 [95%CI 1.25–1.56] in CKD, 1.49 [95%CI 1.37–1.62] in T2D), especially when both were present (2.19 [95%CI.91-2.51]) 19. In the Kaiser Permanente cohort, T2D patients were 1.5-3 times more likely to die from any cause than patients without T2D in all categories of eGFR and for all levels of albuminuria 20.
Although all-cause and cardiovascular mortality is higher in individuals with HTN than in those without 21, the associations of eGFR and ACR with mortality outcomes were found to be stronger in individuals without hypertension than in those with hypertension 21.
Several studies have reported the higher risk of death associated with HTN and T2D in individuals with CKD, either similar (HR 1.25 [95%CI 1.12–1.39] and 1.57 [95%CI 1.29–1.92] respectively, CHS) or higher for T2D than HTN (HR 1.11 [95%CI 1.02–1.21] and 1.61 [95%CI 1.52–1.70] respectively 22.
However, the interaction between HTN and T2D, two frequently coexisting conditions in CKD, has not been fully evaluated.
According to these results, the presence of both T2D and HTN would have a higher risk of death than HTN alone, but T2D had the highest risk. On the contrary, there were no differences between HTN and T2D for CVE, and the risk increased when both were present. Moreover, the risk associated with older age, higher comorbidity, eGFR and albuminuria were higher for death than for CVE, and lower, or non-significant, with non-controlled HbA1c and SBP. Therefore, the potential impact and benefit of control, for the latter risk factors would be greater for CVE.
The known excess mortality risk in T2D has been currently described at about 15% 23,24. When adjusting for SBP and other risk factors, the association strengthened between T2D status and ACM but not for cardiovascular mortality 24.
The results also resembles those of the meta-analysis of risk factors for CVE and death in individuals with eGFR < 30, where both diabetes (HR 1.41; 95%CI 1.30–1.53) and SBP ≥ 140 mmHg (HR 1.09; 95%CI 1.04–1.15) increased the risk of CVE, but only diabetes (HR 1.12; 95%CI 1.03–1.22) increased the risk of death 25.
In another study using electronic health records in individuals with CKD with/without diabetes and no prior CVE 26, the risk of major CVE was 4.6–2.4 times higher in those with T2D. The presence of a HTN diagnosis increased the risk a 10% in Non-Diabetes (HR 1.09; 95%CI 1.03–1.15) but was not significant in T2D (1.12; 0.99–1.27). Moreover, higher BP measures tended to increase the risk more in Non-DM than T2D. HTN is highly prevalent in T2D. In the present analysis, as for T2D, we considered HTN diagnosis and BP control separately, so this effect could appear more clearly, without differences between HTN-CKD and T2D-CKD.
The results of the study might have important clinical significance. Studies show a big room for improvement in CKD screening and management, especially for albuminuria 27, 28.. In the present CKD cohort, albuminuria was assessed in 52% of cases, more frequently when associated with T2D diagnosis. The onset of CKD in the first years after T2D and/or HTN diagnosis can identify individuals at higher risk of adverse events, and screening should be emphasized. Moreover, although the cardiovascular risk is already increased when CKD is present, the coexistence of HTN and T2D identifies individuals at even higher risk, and should prompt a more intensive management to reduce it, including a better control of BP and HbA1c.
Only a small proportion of patients undergo biopsy, which are performed in specific circumstances usually to rule out other causes of renal disease. In a cohort followed by nephrologists, patients with T2D had a similar risk of ESRD when it was considered the primary cause of CKD according to the consulting physician (HR 1.49; 95% CI 1.18–1.88) or only as comorbidity (1.57; 95% CI 1.14–2.15) 29. Even though a causal approximation, T2D and HTN diagnosis are easily identified and can add useful information in the management of CKD patients.
The results of the present study should be interpreted considering some limitations. Some refer to the type of study. Causal relationships can only be approached in these study, we aimed to make descriptions and report associations. Data are obtained from electronic health records which include primary care medical histories of subjects seeking care in the geographic primary care area covered by the Institut Català de la Salut and misdetection cannot be excluded. However, data for cardiovascular disease in primary health care has been shown to be of higher quality than for other diseases and suitable for epidemiological studies in our population 13.
Although models were adjusted for socioeconomic factors, cardiovascular risk factors and diseases, comorbidities and treatments, the presence of unmeasured or residual confounding cannot be ruled out. In this sense, renal disease could act as a marker of multimorbidity.
The eGFR was estimated from serum creatinine measurements using the CKD-EPI formula, and limitations of creatinine-based estimating formulas, with greater impact at higher eGFR, must be accepted. Data on race were not available and correction could not be applied, although Caucasian ethnicity is vastly predominant in the population under study. Although, creatinine was measured in different labs, and different methods may be used, all have completed standardization of creatinine calibration to be traceable to an isotope dilution mass spectrometry (IDMS) reference measurement procedure, which reduces variability 4. Two different measures of albumin in urine were used. The multiple meta-analyses performed by the CKD Prognosis Consortium have also included different methods without modification of the results 1,2,3,21,.
The risk of stroke is higher in women and increases with age 30. Individuals in this primary care incident CKD cohort are more frequently women and quite old and stroke may be overrepresented. Therefore, results might not apply to younger people with different patterns of cardiovascular disease. However, that is the type of population usually attended in primary care and represents the majority of CKD patients.
Despite these limitations, the study presents noteworthy strengths. To our knowledge, it includes the largest incident CKD sample, with diagnosis confirmed using not epidemiological but clinical criteria and with a considerable follow-up. Furthermore, the use of real world data is representative of the usual care. The wealth of information on drug prescription, lab parameters measurements, and complete medical histories is also remarkable. In addition, in Catalonia,, as in other regions of Europe like in the UK or in the rest of Spain, primary care acts as a gate-keeper of patient management. Therefore, it can be considered that a primary care database like SIDIAP provides a quite complete picture of a CKD patient cohort in terms of clinical characterization.