Faye L. Norby

Key Points: Using proteomics and Mendelian randomization, we identified causal proteins for the development of atrial fibrillation in patients with CKD. Immunologic function, cell growth, and neural development were common biologic pathways underlying atrial fibrillation and left atrial enlargement. Machine learning led to a nine-protein risk model that had a similar predictive performance for incident atrial fibrillation as clinical risk scores. Background: CKD is strongly associated with atrial fibrillation. Understanding the biologic pathways for this association and creating predictive models have been challenging. Left atrial enlargement is a substrate for atrial fibrillation, but any overlap between biomarkers of atrial fibrillation and left atrial enlargement in patients with CKD is unknown. Methods: We evaluated 4590 plasma proteins with SomaScan in two cohorts of adult patients with CKD: the Chronic Renal Insufficiency Cohort (CRIC; n =2654) and the Atherosclerosis Risk in Communities cohort (ARIC; n =1326). Using Mendelian randomization, we identified proteins along the causal pathway to atrial fibrillation. We also identified proteins and corresponding pathways associated with larger echocardiographic left atrial size, a recognized substrate for atrial fibrillation. Finally, we developed and validated a multiprotein risk score for incident atrial fibrillation in the CKD population. Results: Over 5 years, incident atrial fibrillation occurred among 150 patients in the CRIC and 140 in the ARIC cohort. We identified three proteins causally linked to incident atrial fibrillation: neural EGF like‑like protein 1, cartilage intermediate layer protein 2, and matrix metallopeptidase 12. Pathway analysis revealed an overlap in eight of the top ten canonical pathways for incident atrial fibrillation and left atrial enlargement. A risk model for incident atrial fibrillation composed of proteins had annualized areas under the receiver-operating characteristic curve over 5 years ranging from 0.65 to 0.76, a performance similar to the Cohorts for Heart and Aging Research in Genomic Epidemiology Atrial Fibrillation (CHARGE-AF) clinical risk score. Conclusions: This study identified causal proteins and biologic mechanisms underlying incident atrial fibrillation in CKD. A proteomic risk score for incident atrial fibrillation in CKD performed similarly to the CHARGE-AF clinical risk score.

Emulating the target trial framework in pharmacoepidemiology is challenging when there is no active comparator. We evaluate six approaches to finding surrogate index dates for untreated patients with the goal of identifying one or more solutions that indicate they would give potentially unbiased results. This numerical experiment used 73 070 patients from the MarketScan administrative databases (2013-2019) with type 2 diabetes, first-line therapy with metformin, and second-line therapy with either sodium-glucose cotransporter 2 inhibitors (SGLT2is) or sulfonylureas. Patients taking sulfonylureas were converted into an experimental “untreated” arm. Part 1 sought to find surrogate index dates for the untreated arm. Part 2 compared the experimental estimates of the effect of SGLT2is on cardiovascular disease compared to sulfonylureas, using the surrogate index dates, to the reference estimate. The reference hazard ratio was 0.69. The hazard ratios after the respective approaches for selecting surrogate index dates are as follows: rejection sampling: 0.61, 0.63; median: 1.10, 1.15; prediction model: 0.96; matching algorithm: 1.07. Only the rejection sampling approaches for selecting a surrogate index date provided results which indicate low amounts of potential bias. Extreme care should be taken when making study design decisions for observational research questions that lack an active comparator group.

Supplemental Table 10 shows that associations between mean of Visit 2 and Visit 3 sMICA and sMICB levels (continuous variables) with CRC risk.

Supplemental Table 6 shows the associations of sMICA and sMICB levels (continuous variables) and pre-selected SNPsb with risk of CRC and subtypes of CRC.

Supplemental Table 7 shows the association of sMICA and sMICB levels (continuous variables) and pre-selected SNPsa with risk of CRC stratified by family history of CRC.

Supplemental Table 3 shows the associations between pre-selected SNPs and CRC risk stratified by sex, race, or diabetes status.

Supplemental Table 4 shows the associations of Visit 2 sMICA and sMICB levels (in quartiles) with CRC risk.

Supplemental Table 8 shows the associations of pre-selected SNPs with Visit 2 sMICA and sMICB levels and CRC risk without adjustment for ancestry markers.

Supplemental table 5 shows the associations between pre-selected SNPs and risk of CRC with and without adjustment for sMICA and sMICB levels.

Supplemental Table 9 shows the associations between pre-selected SNPs and mean of Visit 2 and Visit 3 sMICA and sMICB levels (continuous variables).