Both diabetes and chronic kidney disease are a burden to health spendings. Avoiding complications of CKD such as dialysis is important. A better monitoring of at risk patients is therefore needed. Using risk-score to discriminate between high-risk patients and low-risk patients can be helpful to target the most vulnerable population.

We assessed the difference in drug prescription between high and low risk patients to evolve to terminal chronic kidney disease.

Using the french cohort ND-CRIS a 3-year prognosis score of evolution to terminal chronic kidney disease, was developed using a logistic regression model. The model was validated internally.

The population was then divided between high-risk and low-risk for a risk-threshold of 0,10. Using chi-2 tests, prescription drugs were compared between the two groups, in different time frame (inclusion, T+1 year, T+ 2 years). Time-evolution of prescription was also assessed. Sensitivity analyses were performed using different risk thresholds.

At inclusion, differences were observed concerning prescriptions of anti-diabetes treatment such as insulin (p = 0.012) in high risk group, metformin (p<0,0001), gliclazid (p=0,0065), sitagliptin (p=0,014) in the low risk group. Concerning anti-hypertensive drugs difference exist : calcic canal inhibitors (p=0,037) are more prescribed in high risk patients.

Calcic canal inhibitors prescription differences substist at T+1 (p=0,0151), as well as insulin (p=0,0025), and metformin (p=0,0318). At T+2, the only difference is the prescription of amlodipin (p=0,0003).

In the cohort, prescription of opioids (p=0,002), glinids (p=0,0037), iron (p=0,028), acetaminophen(p=0,0011) are increasing, while metformin, beta-blockers (p=0,0009), diuretic (p < 0,0001) and statins (p=0,029) are decreasing.

These results are confirmed in the sensitivity analyses.

Even unknowingly practitioners are differentiating prescription between the high risk group and the low risk group. Such differences indicate the possibility to personalize patient follow-up with no interference on quality of care.