Multiple versus single and other estimates of baseline proteinuria status as predictors of adverse outcomes in the general population

Abstract

Background: The association of proteinuria and adverse clinical outcomes is well established. The optimal method of classifying proteinuria status for study participants in whom it is measured multiple times is unknown, especially when the frequency of measurement varies between participants.

Study design: Population-based longitudinal study.

Setting & participants: All adults with at least one outpatient serum creatinine measurement in the province of Alberta, Canada.

Factor: Proteinuria (dipstick, albumin-creatinine ratio [ACR]).

Outcomes: All-cause mortality, end-stage renal disease, or doubling of serum creatinine level.

Measures: All outpatient urine dipstick and ACR measurements in the 6-month period before and after the first (index) estimated glomerular filtration rate were used to establish baseline proteinuria. Dipstick measures were analyzed as ceiling (median value up to the next integer), floor (median value down to the next integer), high (single highest dipstick value), low (single lowest dipstick value), and first (first available dipstick value only). Measurements of ACR were evaluated similarly and a median (median of all ACR measurements) value was added.

Results: Of 920,985 participants, 17% (n = 160,548) had multiple dipstick urinalysis measurements and 22% (n = 22,814) had multiple ACR measurements. With single measurements, absolute rates of mortality and renal outcomes were lower in every proteinuria category compared with multiple measurements. In contrast, the relative increase in rate ratio was greater with increasing proteinuria in patients with single measurements compared with those with multiple measurements. In all classification systems evaluated, more severe proteinuria was associated with significantly higher rates of both outcomes (all P for trend <0.001).

Limitations: Lack of a gold standard for choosing between methods.

Conclusions: Rates of adverse outcomes related to multiple baseline proteinuria/albuminuria measurements were similar, independent of the measure of baseline proteinuria that was used to combine results. In contrast, discarding follow-up measurements and relying on only the first measurement led to lower estimates of absolute and relative risk for each proteinuria category.