Hypertension, proteinuria, and RAAS inhibition use in children with systemic lupus erythematosus: Data from a multi-institutional pediatric learning health system

Abstract

ObjectivesTo assess the potential of a multi-institutional pediatric learning health system for comparative effectiveness research in pediatric-onset systemic lupus erythematosus (SLE), we characterized renin angiotensin aldosterone system (RAAS) inhibitor utilization and the feasibility of ascertaining key treatment indications and outcomes, including hypertension and proteinuria.MethodsWe identified children with SLE and lupus nephritis (LN) at 6 PEDSnet institutions using previously developed computable phenotypes. A reference population of controls without SLE was randomly sampled from the same rheumatology/nephrology clinics (N = 200 controls per site). We evaluated data completeness, plausibility, and conformance related to RAAS inhibitor treatment indications and outcomes: (a) percent of encounters with blood pressure (BP) readings, (b) percent of BP readings with paired height, (c) percent of patients with ≥1 urinalysis within 7 days of SLE diagnosis, (d) urinalyses for which proteinuria could be classified as normal or abnormal, and (e) RAAS inhibitor use.ResultsThere were 1303 patients with SLE, including 457 with LN. BP measurements were available at 62% of encounters, of which 96% were paired with a height within 90 days. BP distributions were higher in patients with SLE and LN compared to controls without SLE. One third of SLE patients had a hypertension diagnosis. 60%-83% of patients at each site had a urinalysis protein measurement within 7 days of SLE diagnosis. A normal or abnormal result for proteinuria could be derived for 96% of measurements, and nearly all patients with LN had ≥1 abnormal result (range 94%-100% by site). RAAS inhibitors were prescribed to 31% of SLE and 71% of LN patients (range 60%-84% by site). Hypertension, elevated BP or proteinuria was identified in 90% of SLE cases at the time of RAAS inhibitor initiation.ConclusionWe demonstrated the feasibility of ascertaining health measurement data relevant to pediatric lupus treatment and outcomes in a pediatric learning health system, as well as hospital-level variation in RAAS inhibitor use. This work will facilitate more efficient multi-institutional comparative effectiveness research and also highlights opportunities for increased treatment standardization.

Keywords: Systemic lupus erythematosus; cardiovascular disease; renal lupus.

Figure 1.

Distribution of systolic blood pressure (SBP) mmHg values in the A) SLE and B) LN cohorts (pink) as compared to the reference population of non-SLE controls followed in the same rheumatology and nephrology clinics (blue), stratified by site of care. Distributions include all available SBP measurements during follow-up for each patient at each of the six PEDSnet sites. Compared to non-SLE controls, there were trends towards higher SBP values in patients with SLE +/− LN. The site-level mean SBP ranged from 116 to 125 among all SLE patients across sites, and 119 to 133 among all LN patients. The analysis was not intended to compare SBP between sites.

Figure 2.

Distribution of diastolic blood pressure (DBP) mmHg values in the A) SLE and B) LN cohorts (pink) as compared to the reference population of non-SLE controls followed in the same rheumatology and nephrology clinics (blue), stratified by site of care. Distributions include all available DBP measurements during follow-up for each patient at each of the 6 sites. Compared to non-SLE controls, SLE patients at 5 of 6 sites tended towards higher DBP values. Patients with LN tended towards higher DBP at all sites. Site-level mean DBP ranged from 67–72 and 70–75 for all SLE and LN patients, respectively. The analysis was not intended to compare DBP between sites.

Figure 3.

Proportion of urinalysis protein measurements with an evaluable derived result for the A) total SLE cohort and B) LN subgroup, stratified by site of care.