Review The Utility of 12 Hour Urine Collection for the Diagnosis of Preeclampsia A Systematic Review and Meta analysis

stout2015 pdf Review The Utility of 12 Hour Urine Collection for the Diagnosis of Preeclampsia A Systematic Review and Meta analysis Molly J Stout, MD, MSCI, Shayna N Conner, MD, MSCI, Graham A Coldit.

The Utility of 12-Hour Urine Collection for the Diagnosis of Preeclampsia

A Systematic Review and Meta-analysis

Molly J Stout,MD,MSCI, Shayna N Conner, MD,MSCI, Graham A Colditz, MD,DrPH,

George A Macones,MD,MSCE, and Methodius G Tuuli, MD,MPH

OBJECTIVE: To systematically review the literature and

synthesize data on the diagnostic performance of a

12-hour urine collection for proteinuria in pregnant women

with suspected preeclampsia.

DATA SOURCES: We performed a literature search of

PubMed, Embase, Scopus, ClinicalTrials.gov, and CINAHL

through February 2014 using key words related to

gesta-tional hypertension, preeclampsia, and proteinuria.

METHODS OF STUDY SELECTION: Studies that

con-tained results of both the 12-hour and 24-hour urine

collection in the same patients were eligible.

TABULATION, INTEGRATION, AND RESULTS: Three

independent reviewers abstracted test performance

characteristics from each study for the performance of

a 12-hour urine collection for the diagnosis of

pro-teinuria defined as 300 mg in 24 hours Diagnostic

meta-analysis was performed to obtain summary

statis-tics Heterogeneity was assessed using the Cochrane Q

or I 2 Receiver operating characteristic curve analysis was

used to assess the optimal diagnostic cutpoint for

pro-teinuria from a 12-hour urine collection Stratified

anal-ysis was performed based on whether patients were on

bed rest during urine collection A total of seven studies

met inclusion criteria The 12-hour urine protein was

overall highly predictive of proteinuria on 24-hour urine collection area under receiver operating characteristic curve: 0.97 (95% confidence interval [CI] 0.95–0.98) The pooled sensitivity was 92% (95% CI 86–96) and spec-ificity was 99% (95% CI 75–100) The optimal cutpoint based on the receiver operating characteristic curve was 150 mg of protein on 12-hour collection.

CONCLUSION: A 12-hour urine collection compares favorably with a 24-hour urine collection for the diag-nosis of proteinuria in women with suspected pre-eclampsia and has the advantage of convenience and improved clinical efficiency.

(Obstet Gynecol 2015;126:731–6) DOI: 10.1097/AOG.0000000000001042

Preeclampsia complicates 5–8% of pregnancies, is

characterized by hypertension and other end-organ injury, and remains a leading cause of maternal morbidity and mortality in the United States.1,2 A recent American College of Obstetricians and Gyne-cologists Task Force on hypertension in pregnancy removed the requirement of proteinuria for the diag-nosis of preeclampsia if there are other findings sug-gestive of end organ involvement (thrombocytopenia, elevated liver transaminases, renal insufficiency, pul-monary edema, or new-onset neurologic symptoms).2

However, in the absence of these severe features, quantification of urinary protein remains an impor-tant diagnostic step for evaluation of hypertension

in pregnancy

Urine protein can be quantified using either a 24-hour urine collection or a spot urine protein-to-creatinine ratio.2 Several studies have investigated urine protein-to-creatinine ratio as a rapid test to obvi-ate the need for a 24-hour urine collection.3–12

Although the data show that extremely high or low urine protein-to-creatinine ratio values may be a sub-stitute for a 24-hour urine collection,6,13 there are

From the Departments of Obstetrics and Gynecology and Surgery, Washington

University School of Medicine in St Louis, St Louis, Missouri.

Dr Stout and Dr Tuuli are supported by a Women’s Reproductive Health

Research Career Development grant from the Eunice Kennedy Shriver

National Institute of Child Health and Human Development (National

Insti-tutes of Health/Eunice Kennedy Shriver National Institute of Child Health

and Human Development—1K12HD063086-01).

Corresponding author: Molly J Stout, MD, MSCI, Washington University

School of Medicine, Campus Box 8064, 4566 Scott Avenue, St Louis, MO

63110; e-mail: stoutm@wudosis.wustl.edu.

Financial Disclosure

The authors did not report any potential conflicts of interest.

© 2015 by The American College of Obstetricians and Gynecologists Published

by Wolters Kluwer Health, Inc All rights reserved.

ISSN: 0029-7844/15

clinical circumstances that may still require a 24-hour

urine collection

Several studies have investigated the use of a

12-hour urine collection as opposed to a 24-12-hour urine

collection for the diagnosis of proteinuria.14–20 We

conducted a systematic review for relevant studies

evaluating the diagnostic utility of a 12-hour urine

collection compared with a 24-hour urine collection

for the diagnosis of proteinuria in pregnant women

with suspected preeclampsia We then performed

a diagnostic meta-analysis of the data to obtain

sum-mary diagnostic characteristics and estimate the

opti-mal cutpoint of protein on 12-hour urine collection for

the diagnosis of proteinuria

SOURCES

We used a predesigned study protocol outlining the

search strategy, study inclusion and exclusion criteria,

quality assessment tool, and data analysis plan The

protocol adhered to guidelines in the Cochrane

Handbook for Systematic Reviews of Diagnostic

Accuracy.21We searched PubMed, Embase, Scopus,

ClinicalTrials.gov, and CINAHL through February

2014 using the MeSH: “hypertension,

pregnancy-Induced,” “gestational hypertension,” “pregnancy

tran-sient hypertension,” “edema-proteinuria-hypertension

gestosis,” “pregnancy toxemia,” “pre-eclampsia,”

“proteinuria,” “albuminuria,” “protein,” “urinary

protein,” “12-hour,” “12-h,” and “12-hours.” Two

medical librarians with specific training in literature

searches for systematic reviews led the search

STUDY SELECTION

Studies that contained results of both the 12-hour and

24-hour urine collection in the same patients were

eligible We excluded unpublished studies,

non-English studies, conference proceedings, abstracts, case

studies, and commentaries Two independent

investi-gators (M.J.S and S.N.C.) filtered titles and abstracts

identified from the initial search to determine whether

the studies met inclusion criteria Citation lists from

included articles were manually searched Once

screen-ing of titles and abstracts was performed, three

independent reviewers (M.J.S., S.N.C., M.G.T.) read

the article and recorded quality data, incidence of

positive 24-hour urine protein (greater than 300 mg),

cutpoint used to define positive 12-hour urine protein,

true-positives and -negatives, and false-positives and

-negatives To be included in the final analysis, the

study must have been performed in pregnant patients

at or beyond 20 weeks of gestation, included

collec-tions of both a 12-hour and a 24-hour urine specimen,

reported total protein (not concentration), and

provided enough diagnostic information (prevalence

of proteinuria in sample, cutpoint used for 12-hour test, sensitivity, specificity, predictive values) that true-positives and -negatives, and false-true-positives and -neg-atives could be back-calculated The true-positives and -negatives and false-positives and -negatives were based

on the cutpoint used in the individual studies Disagree-ments were resolved by consensus of all three reviewers

We assessed study quality based on the QUADAS tool for diagnostic meta-analysis The QUADAS tool was designed specifically to assess bias and study quality specifically related to studies on diagnostics tests and includes yes or no questions such as whether the spectrum of patients who underwent the test is representative of those who would have the clinical disease, whether the reference test and the index test were interpreted blindly, whether enough information

is provided about the reference and index tests to replicate them, and other questions specifically perti-nent to diagnostic test studies All reviewers extracted quality data based on the QUADAS tool.22 One QUADAS question (“Was the reference standard inde-pendent of the index test?”) was not pertinent to this study and was not assessed

Results of the 12-hour urine results were com-pared with those of the gold standard 24-hour urine for each individual study to create two-by-two tables containing true- and false-positive and -negatives Meta-analysis was used to calculate pooled sensitivity and specificity of 12-hour urine compared with 24-hour urine for quantification of proteinuria

We used the Dersimonian and Laird random-effects models to pool data irrespective of demonstra-ble statistical heterogeneity Heterogeneity was

Fig 1 Flowchart of studies examined for inclusion in the meta-analysis.

Stout 12-Hour Urine Protein Obstet Gynecol 2015.

assessed using the Higgin’s I2statistic and Cochrane’s

Q test Stratified analysis according to bed rest status

during the urine collection was performed

A summary receiver operating characteristic

(ROC) curve was used to estimate the optimal

cutpoint for the 12-hour urine protein that maximized

both sensitivity and specificity The optimal cutpoint

on the ROC curve was identified as the left uppermost

point The 12-hour protein cutpoint used in the study

closest to this point was considered the optimal

cutpoint

Publication bias was assessed using Deek’s funnel

plot for small study effect This is a regression plot of

the diagnostic log odds ratio against the inverse of the

sample size weighted by the effective sample size.23

P,.10 was considered significant to take into account

the modest statistical power of this test

Analysis was performed using the MIDAS

pack-age in Stata12

RESULTS

A total of 87 studies were identified from the initial

search, of which seven were included in the final

analysis (Fig 1) The incidence of 24-hour urine

pro-tein greater than 300 mg ranged from 14% to 86%

The cutpoint used for a positive 12-hour urine

collec-tion ranged from 100 to 165 mg Studies varied with

respect to inpatient or outpatient status as well as the

use of bed rest or not during the urine collection Of

the 410 total number of patients in the studies

included, five (1.2%) had 2-hour urine protein test

results that were false-positive and 16 (3.9%) had

12-hour urine test results that were false-negative (Table 1)

The proportion of studies complying with each of the criteria of the QUADAS tool is shown graphically

in Figure 2 All included studies were a prospective cohort design, had both 12- and 24-hour analysis per-formed by same laboratory, and included patients who by definition underwent both the index (12-hour urine protein) and referent (24-(12-hour urine pro-tein) tests All studies included patients who represent

Table 1 Characteristics of Studies Included in the Meta-analysis Comparing 12-Hour Urine Protein With

24-Hour Urine Protein for the Diagnosis of Preeclampsia

Study

Study

(n)

Incidence of 24-h Greater Than 300

mg (%)

Cutpoint Used for 12-h Urine

U.S.

Location Inpatient

Bed Rest or Modified Bed Rest

True-Positive

True-Negative

False-Positive

False-Negative Rinehart

et al, 15 1999

Adelberg

et al, 16 2001

Schubert and

Abernathy, 17

2006

Moslemizadeh

et al, 19 2008

Tun et al, 14

2012

Rani Singhal

et al, 20 2014

Fig 2 Bar chart showing quality assessment using quality assessment of diagnostic accuracy studies criteria.

Stout 12-Hour Urine Protein Obstet Gynecol 2015.

those who would likely be undergoing testing in

typ-ical clintyp-ical practice (eg, hypertensive disease beyond

20 weeks of gestation) and had clinical data available

that would routinely be available in clinical practice

Studies differed as to whether the 12- and 24-hour

urines were interpreted blindly

The pooled sensitivity was 92% (95% confidence

interval [CI] 86–96%), and specificity was 99% (95%

CI 75–100%) (Fig 3) There was significant

heteroge-neity between studies (I2553%, Q test P,.04 and

I2575%, Q test P,.01, respectively) A planned

strat-ified analysis according to bed rest status was

per-formed because clinical characteristics such as bed

rest compared with ambulation may alter protein

excretion.24,25 Bed rest during the urine collection

was associated with a nominally lower sensitivity

(88% compared with 94%) and slightly higher

speci-ficity (100% compared with 86%) (Table 2) However,

overlapping CIs suggest a nonstatistical difference

The summary ROC curve (comparing sensitivity

on the Y-axis and 1-specificity [false-positives] on the

X-axis) showed that the 12-hour urine protein was

highly predictive of proteinuria on 24-hour urine

collection (area under ROC curve 0.97, 95% CI 0.95–0.98; Fig 4) The 12-hour protein cutpoint used

in the study closest to the optimal cutpoint (located in the upper left most region of the ROC curve where sensitivity is maximized and false-positives are minimized) was 150 mg.17

The Deek’s funnel plot showed no significant cor-relation between the diagnostic odds ratio and effec-tive sample size, suggesting no significant publication bias (Fig 5)

DISCUSSION

The findings of this systematic review and diagnostic meta-analysis suggest that a 12-hour urine protein collection performs well compared with a 24-hour urine collection for the diagnosis of preeclampsia with high sensitivity and specificity

Hypertensive disease continues to be a major source of pregnancy-related morbidity The differen-tiation of preeclampsia from gestational hypertension depends in part on quantification of proteinuria A recent American College of Obstetricians and Gyne-cologists Task Force document recommends against the use of dipstick quantification of urine protein.2

Other rapid quantification methodologies such as urine protein-to-creatinine ratio are useful in certain circumstances, but recent meta-analyses suggest that urine protein-to-creatinine ratio may have most utility

in ruling out (rather than ruling in) significant protein-uria.6,13 Thus, 24-hour urine collection for protein is still clinically relevant and required for some patients This analysis has several strengths including use of

a predefined protocol adhering to guidelines for meta-analysis of diagnostic studies,21 the comprehensive

Table 2 Pooled Sensitivity and Specificity by Bed

Rest Subgroup

Subgroup

Sensitivity (95% CI)

Specificity (95% CI) Bed rest (n54 studies, 220

patients)

88 (80–96) 100 (NA)

No bed rest (n52 studies,

65 patients)

94 (88–100) 86 (67–100)

CI, confidence interval; NA, not applicable.

Fig 3 Forest plot of pooled sensitivity (A) and specificity (B).

Stout 12-Hour Urine Protein Obstet Gynecol 2015.

literature search performed by trained medical librarians,

and use of the ROC curve to objectively estimate the

optimal cutpoint The study also has a number of

limi-tations We included only seven studies with relatively

small samples sizes that met inclusion criteria Moreover,

there was also significant between study heterogeneity

likely resulting from the varied cutpoints used and

differ-ences in the use of bed rest during urine collection We

used the random-effects model to account for this

het-erogeneity and obtain conservative estimates of the

diag-nostic characteristics In addition, because of the small

number of studies in subgroups, it is unclear whether

urine collection during the day or night and bed rest

or no bed rest alter the prediction of proteinuria

In conclusion, results of this systematic review and

diagnostic meta-analysis suggest that a 12-hour urine

collection performs well for the diagnosis of

pro-teinuria in hypertensive women during pregnancy A

clinically applicable cutpoint based on the data

avail-able would be 150 mg per 12-hour collection This cutpoint is associated with 99% specificity and 92% sensitivity Use of the 12-hour urine collection would

be more convenient and expedite diagnosis, clinical management, and decrease cost Future studies should overcome the limitations of the studies included in this meta-analysis by using larger sample sizes, collecting urine in a homogenous manner, and stratifying the 12-hour urine collection by day or night

REFERENCES

1 Callaghan WM Overview of maternal mortality in the United States Semin Perinatol 2012;36:2–6.

2 American College of Obstetricians and Gynecologists; Task Force on Hypertension in Pregnancy Hypertension in preg-nancy Report of the American College of Obstetricians and Gynecologists’ Task Force on Hypertension in Pregnancy Obstet Gynecol 2013;122:1122–31.

3 Stout MJ, Scifres CM, Stamilio DM Diagnostic utility of urine protein-to-creatinine ratio for identifying proteinuria in preg-nancy J Matern Fetal Neonatal Med 2013;26:66–70.

4 Durnwald C, Mercer B A prospective comparison of total pro-tein/creatinine ratio versus 24-hour urine protein in women with suspected preeclampsia Am J Obstet Gynecol 2003;189:848–52.

5 Papanna R, Mann LK, Kouides RW, Glantz JC Protein/crea-tinine ratio in preeclampsia: a systematic review Obstet Gyne-col 2008;112:135–44.

Diagnostic odds ratio

.10

1

.15

.20

.25

Singhal 2014 21

Tun 2012 15

Moslemizadeh 2008 20

Rabiee 2006 19

Schubert 2006 18

Adelberg 2001 17

Rinehart 1999 16

Regression line

Fig 5 Deek’s funnel plot for publication bias The colored circles represent the individual studies The regression line represents the relationship between the inverse of the effective sample size and the effect estimate (diagnostic odds ratio) The nonsignificant correlation suggests absence

of publication bias, P5.23.

Stout 12-Hour Urine Protein Obstet Gynecol 2015.

1-specifi city

1.0

0.5

0.0

Singhal 2014 21

Tun 2012 15

Moslemizadeh 2008 20

Rabiee 2006 19

Schubert 2006 18

Adelberg 2001 17

Rinehart 1999 16

SROC curve

1.0 0.5

0.0

Summary operating point 95% confi dence contour 95% prediction contour

Fig 4 Summary receiver operating characteristic (SROC)

curve for summary sensitivity and specificity for diagnosis

of proteinuria by 12-hour urine protein The colored circles

represent the individual studies Summary operating point:

sensitivity50.92 (95% confidence interval [CI] 0.86–0.96),

specificity50.99 (95% CI 0.75–1.00) Summary receiver

operating characteristic curve: area under the curve50.97

(95% CI 0.95–0.98).

Stout 12-Hour Urine Protein Obstet Gynecol 2015.

6 Sanchez-Ramos L, Gillen G, Zamora J, Stenyakina A,

Kaunitz AM The protein-to-creatinine ratio for the prediction

of significant proteinuria in patients at risk for preeclampsia:

a meta-analysis Ann Clin Lab Sci 2013;43:211–20.

7 Verdonk K, Niemeijer IC, Hop WC, de Rijke YB, Steegers EA,

van den Meiracker AH, et al Variation of urinary protein to

creatinine ratio during the day in women with suspected

pre-eclampsia BJOG 2014;121:1660–5.

8 Al RA, Baykal C, Karacay O, Geyik PO, Altun S, Dolen I.

Random urine protein-creatinine ratio to predict proteinuria

in new-onset mild hypertension in late pregnancy Obstet

Gy-necol 2004;104:367–71.

9 Neithardt AB, Dooley SL, Borensztajn J Prediction of 24-hour

protein excretion in pregnancy with a single voided urine

protein-to-creatinine ratio Am J Obstet Gynecol 2002;186:883–6.

10 Robert M, Sepandj F, Liston RM, Dooley KC Random

protein-creatinine ratio for the quantitation of proteinuria in

pregnancy Obstet Gynecol 1997;90:893–5.

11 Rodriguez-Thompson D, Lieberman ES Use of a random

uri-nary protein-to-creatinine ratio for the diagnosis of significant

proteinuria during pregnancy Am J Obstet Gynecol 2001;185:

808–11.

12 Saudan PJ, Brown MA, Farrell T, Shaw L Improved methods

of assessing proteinuria in hypertensive pregnancy Br J Obstet

Gynaecol 1997;104:1159–64.

13 Côté AM, Brown MA, Lam E, von Dadelszen P, Firoz T,

Liston RM, et al Diagnostic accuracy of urinary spot protein:

creatinine ratio for proteinuria in hypertensive pregnant

women: systematic review BMJ 2008;336:1003–6.

14 Tun C, Quiñones JN, Kurt A, Smulian JC, Rochon M

Com-parison of 12-hour urine protein and protein:creatinine ratio

with 24-hour urine protein for the diagnosis of preeclampsia.

Am J Obstet Gynecol 2012;207:233.e1–8.

15 Rinehart BK, Terrone DA, Larmon JE, Perry KG Jr, Martin RW,

Martin JN Jr A 12-hour urine collection accurately assesses

pro-teinuria in the hospitalized hypertensive gravida J Perinatol 1999;19:556–8.

16 Adelberg AM, Miller J, Doerzbacher M, Lambers DS Corre-lation of quantitative protein measurements in 8-, 12-, and 24-hour urine samples for the diagnosis of preeclampsia Am J Obstet Gynecol 2001;185:804–7.

17 Schubert FP, Abernathy MP Alternate evaluations of proteinuria

in the gravid hypertensive patient J Reprod Med 2006;51:709–14.

18 Rabiee S Comparison of predictive value of 8, 12, and 24-hour proteinura in pre-eclampsia Pak J Med Sci 2006;23:3.

19 Moslemizadeh N, Yousefnejad K, Moghadam TG, Peyvandi S Urinary protein assessment in preeclampsia: which sample is more suitable? Pak J Biol Sci 2008;11:2584–8.

20 Rani Singhal S, Ghalaut V, Lata S, Madaan H, Kadian V, Sachdeva A Correlation of 2 hour, 4 hour, 8 hour and 12 hour urine protein with 24 hour urinary protein in preeclampsia.

J Family Reprod Health 2014;8:131–4.

21 Macaskill P, Gatsonis C, Deeks J, Harbord R, Takwoingi Y Chapter 10: analysing and presenting results London (UK): The Cochrane Collaboration; 2010.

22 Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J The development of QUADAS: a tool for the quality assess-ment of studies of diagnostic accuracy included in systematic reviews BMC Med Res Methodol 2003;3:25.

23 Deeks JJ, Macaskill P, Irwig L The performance of tests of publication bias and other sample size effects in systematic re-views of diagnostic test accuracy was assessed J Clin Epidemiol 2005;58:882–93.

24 Wan LL, Yano S, Hiromura K, Tsukada Y, Tomono S, Kawazu S Effects of posture on creatinine clearance and uri-nary protein excretion in patients with various renal diseases Clin Nephrol 1995;43:312–7.

25 Deeb A, Hadj-Aissa A, Ducher M, Chapuis-Cellier C, Fauvel JP The best way to detect elevated albuminuria Neph-ron Clin Pract 2011;117:c333–40.