Preeclampsia is one of the leading causes of perinatal morbidity and mortality worldwide.1 In the United States, although maternal mortality has dropped, preeclampsia still causes 18% of the approximately 350 maternal deaths each year.2 Further, it is also a leading cause of indicated preterm delivery, subsequently leading to increased neonatal morbidity and mortality.3 The diagnosis of preeclampsia currently relies on at least two blood pressures above 140/90 6 h apart and proteinuria greater than 300 mg in a 24-h urine collection.4 However, there are times when a decision regarding when to deliver, or whether to utilize magnesium sulfate for seizure prophylaxis,5 needs to be made sooner than 24 h.

In such cases, urine dipstick protein or, more recently, the urine protein-to-creatinine ratio has been utilized.6, 7 Historically, the concern with the standard urine dipstick was the accuracy of this one-time measurement. In particular, the urine dipstick has been demonstrated to have a poor sensitivity and, in studies, greater than 50% of women with negative or trace urine dipsticks have been found to have significant proteinuria.6, 8 Although the urine protein/creatinine ratio should alleviate the problem with overall urine concentration, it too has not been found to have a particularly high sensitivity when a ratio of 0.3 has been used as a diagnostic threshold. Lower ratios of 0.159 and 0.1910 have been suggested, but these continue to demonstrate problems with sensitivity, specificity and the corresponding positive and negative predictive values.

In the current edition of the Journal of Perinatology, Dwyer et al.11 present a potential algorithm for the efficient diagnosis of significant proteinuria. This current algorithm has the strength of recognizing the potential strengths of urine dipstick and urine protein-to-creatinine ratio as having diagnostic thresholds as well as the potential of a reasonable screening threshold. In this case, diagnostic threshold refers to a numerical threshold above which the positive predictive value is essentially 100%. Alternatively, a potential screening threshold should have a sensitivity close to 100%. As Dwyer et al. note, the urine dipstick has a high specificity, but relatively poor sensitivity, but the urine protein-to-creatinine ratio has a high sensitivity in its lower ranges. Their resulting algorithm utilizes these simple steps: (1) all women with a urine dipstick of 1+ or greater are considered to have proteinuria >300 mg in 24 h; (2) all women with a urine dipstick of negative or trace with a urine protein-to-creatinine ratio of 0.28 or greater are considered to have proteinuria >300 mg in 24 h; (3) all women with a urine protein-to-creatinine ratio <0.15 are considered to be negative for significant proteinuria; and (4) women with a urine protein-to-creatinine ratio between 0.15 and 0.27 need to have a 24 h urine collection for formal diagnosis.

As noted above, this relatively simple algorithm notes and uses the strengths and weaknesses of the two simple, one-time urine tests to reliably predict who does and who does not have significant proteinuria and will do so in three out of four patients. The remaining 25% of women will need to undergo the 24-h urine protein collection, and the clinician can utilize other signs and symptoms of preeclampsia to adjust the a priori risk of intervening in the pregnancy or utilizing magnesium sulfate for seizure prophylaxis.

This algorithm deserves future investigation in a large, prospective, multi-center study. Although it is unlikely that in a larger study the thresholds described will lead to 100% positive predictive values and 100% sensitivity, one would hope that both values would be in the 96 to 99% range. Meanwhile, because these data have been replicated in other smaller studies, it seems reasonable to utilize these thresholds to guide acute care of women with elevated blood pressures.