Addressing the problem of inaccuracy of measured 24‐hour urine collections due to incomplete collection

Abstract

The 24‐hour urine collection is widely considered the gold standard for assessing 24‐hour excretion of various analytes. Yet, studies show that >30% of collections are incomplete and hence understate the true 24‐hour excretion. We previously validated an equation for estimating an individual's 24‐hour creatinine excretion, based on weight, sex, race, and age. The present study examines whether routinely correcting measured 24‐hour excretion of an analyte using the ratio of estimated to measured 24‐hour urine creatinine mitigates the misleadingly low values that result from undercollection. Ninety‐nine participants, recruited from an outpatient hypertension clinic, submitted a 24‐hour urine collection for measurement of creatinine and sodium excretion. The urine collections were dichotomized into complete (n = 63) and incomplete (n = 36) collections based on previously described criteria for expected 24‐hour creatinine excretion. To adjust for possible incompleteness of collections, the measured 24‐hour values were then corrected by multiplying them by the ratio of estimated to measured 24‐hour urine creatinine. The mean 24‐hour creatinine excretion was 1682 mg/d among complete collectors. Among incomplete collectors, the mean was 1211 mg/d before and 1695 mg/d after, adjustment. Similarly, mean 24‐hour sodium excretion was 156 mEq/d among complete collectors, whereas among incomplete collectors it was 121 mEq/d before and 171 mEq/d after, adjustment. Undercollection of 24‐hour urines is a common problem and results in misleadingly low values for excretion of analytes being measured. Routine adjustment of 24‐hour urine values using estimated values for 24‐hour creatinine excretion can greatly mitigate artifactually low 24‐hour excretion results.

1. INTRODUCTION

The 24‐hour urine collection is currently the gold standard in measuring the excretion of various analytes or hormones. However, it is fraught with problems. It is very inconvenient and burdensome. Because of this, 24‐hour collections are not suitable for measuring 24‐hour excretion on multiple days, which would be advantageous in assessing excretion of analytes whose excretion varies considerably from day‐to‐day, or for serial monitoring to assess the response to an intervention.1, 2

Another important and common problem is inaccuracy of 24‐hour collections due to undercollection. Completeness of a 24‐hour urine collection is typically assessed from its creatinine content. Studies have shown that 25%‐30% or more of 24‐hour urine collections are undercollected, yielding results that understate the actual 24‐hour excretion of the analyte being assessed.3, 4, 5 In a previous study, we observed that 50% of collections were undercollected, defined as containing <15 mg/kg of creatinine in women and <20 mg/kg in men.3

To address this problem, we developed an equation derived from the easily obtainable subject variables of sex, weight, race, and age to estimate an individual's expected 24‐hour creatinine excretion.3 In developing the equation, the data of participants who provided incomplete 24‐hour collections, as previously defined by Mattix et al,6 were excluded. The accuracy of the derived equation in predicting 24‐hour creatinine excretion was then demonstrated in a validation cohort by the strong correlation observed between the estimated and measured 24‐hour creatinine (r = .80, P < .0001), a correlation that was even stronger (r = .93, P<.0001) among participants whose urine collection was considered complete.3

Importantly, in assessing the accuracy of this equation, differences between estimated and measured creatinine values should not be assumed to be entirely due to inaccuracy of the estimated value; they are to a considerable extent also attributable to inaccuracy in the “gold standard” 24‐hour urine collection itself because of factors such as incomplete collection as well as the considerable day‐to‐day variability of 24‐hour creatinine excretion due to factors such as variation in protein intake and intense exercise.7, 8

Because inaccuracy of the “gold standard” 24‐hour collection artifactually understates the accuracy of our method or of any other alternative method of estimating analyte excretion, we also assessed the accuracy of our equation by comparing the correlation of estimated vs measured 24‐hour creatinine excretion with an alternative gold standard that is independent of the 24‐hour urine collection. To do so, we compared their correlation with muscle mass, which was assessed by bioelectrical impedance. This alternative gold standard was selected because excreted creatinine is a degradation product of muscle and correlates strongly with muscle mass.9, 10, 11, 12 The equation‐derived estimated 24‐hour creatinine excretion was found to correlate extremely well with muscle mass and correlated at least as well as did the measured creatinine excretion in the entire cohort (r = .95 for estimated vs .82 for measured creatinine excretion) and among incomplete collectors (r = .93 for estimated and .85 for measured creatinine excretion).3

These results raised the question of whether routine adjustment of the measured 24‐hour creatinine excretion using the ratio of estimated to measured 24‐hour creatinine excretion should be used to mitigate the widespread problem of inaccuracy of 24‐hour urine collections due to undercollection. The current study was performed to determine the effect of adjusting measured 24‐hour values for sodium and creatinine excretion using the ratio of the estimated to measured creatinine excretion.

2. METHODS

The research protocol was approved by the Institutional Review Board of Weill Cornell Medical College. Written informed consent was obtained from all individual participants included in the study. Ninety‐nine participants were recruited between January 2007 and December 2008. Individuals older than 21 years were recruited irrespective of hypertensive status and renal function, provided that renal function and appetite were stable.

Participants submitted a 24‐hour urine collection, discarding the first morning urine and then collecting all urine up to and including the first urine the next morning. The volume of the 24‐hour urine was recorded, and an aliquot was sent for measurement of sodium and creatinine concentrations by standard laboratory procedures.

The measured 24‐hour creatinine excretion was calculated in standard fashion from the urine volume and creatinine concentration. The estimated 24‐hour creatinine excretion was determined from the following equation, as discussed above3:

$$\text{Estimated}24\text{-}\text{hour}\text{creatinine}\text{excretion}(\text{mg}/24\mathrm{h})=699-421.9(\text{if female})+(7.64\times \text{weight})(\text{pounds})-25.82(\text{if White})-(2.67\times \text{age(years)})$$

In similar fashion, the measured 24‐hour sodium excretion was calculated from the urine sodium concentration and volume. The corrected 24‐hour sodium excretion was then calculated by multiplying the measured value by the ratio of the estimated to measured 24‐hour creatinine excretion. For example, if a participant's estimated 24‐hour creatinine excretion was 2000 mg, and the measured value was only 1000 mg, that is half that value, then the corrected 24‐hour sodium excretion was calculated by multiplying the measured sodium value by two (estimated/measured = 2000/1000 = 2).

For purposes of analysis, the cohort was stratified by completeness of collection into complete collectors (n = 63) and incomplete collectors (n = 39), with the latter defined by 24‐hour urine creatinine excretion <20 mg/kg in men and <15 mg/kg in women.6

2.1. Statistical analysis

Descriptive statistics were used to summarize participant characteristics with means and standard deviations for continuous variables and percentages for categorical variables. Independent samples t tests were used to compare means of estimated and measured values between complete and incomplete collectors. Paired samples t tests were used to compare means of estimated and measured creatinine excretion within complete collectors and within incomplete collectors. Means of measured 24‐hour uncorrected and corrected sodium excretion were similarly compared. Additionally, means between participants who were taking a diuretic were compared with those not taking a diuretic. Correlations between measured and estimated 24‐hour creatinine excretion and between uncorrected and corrected sodium excretion were assessed by Pearson's correlation coefficient. Two‐tailed probability levels for statistical significance tests are reported, with P<.05 considered statistically significant.

Bland‐Altman plots were constructed to evaluate the agreement between the measured 24‐hour urinary creatinine and the estimated 24‐hour urinary creatinine, and between the measured uncorrected sodium excretion and the measured corrected sodium excretion.13 Bias was assessed as the mean individual differences between the measured and estimated 24‐hour creatinine or uncorrected and corrected sodium excretions against the individual mean of these parameters. The individual means of measured and estimated 24‐hour creatinine or measured uncorrected and measured corrected sodium excretion are plotted on the x‐axis, and the individual differences between the two measures are plotted on the y‐axis. The 95% limits of agreement were calculated as within the 1.96 standard deviation of the mean difference. Bland‐Altman plots were constructed separately for complete and incomplete collectors.

3. RESULTS

Selected characteristics of the study participants are shown in Table 1. The majority (82.8%) of participants had hypertension, of whom 76.8% were on treatment, including 60.6% who were taking a thiazide and/or a potassium‐sparing diuretic. Of the 99 participants, 63 met the definition for complete collectors and 36 for incomplete collectors.

Table 1.

Selected characteristics of the study participants (n = 99)

Characteristic	Mean (SD)
Age (y)	61.4 (15.2)
Body mass index (kg/m2)*	27.6 (5.4)
Weight (kg)
Male	88.2 (15.9)
Female	73.3 (18.1)
	%
Male	54.5
Race/Ethnicity
White	82.8
Black	17.2
Hypertensive	82.8
On antihypertensive medication	76.8
Taking a diuretic†	60.6

^*n = 93.

^†Individuals taking a loop diuretic were excluded.

Mean 24‐hour volume, sodium and creatinine excretion for the entire cohort and for the subgroups of complete and incomplete collectors are shown in Table 2. As expected, the measured 24‐hour creatinine and sodium excretion values were lower among incomplete collectors than among complete collectors (P < .0001 and P < .02, respectively). As seen in Table 2, when measured 24‐hour excretion of creatinine and sodium were corrected using the ratio of the estimated to measured 24‐hour creatinine excretion, those differences were no longer observed.

Table 2.

Estimated and measured 24‐h urinary creatinine and sodium excretion

	Entire sample N = 99		Complete collectors N = 63		Incomplete collectors N = 36
	Mean (SD)	Min, Max	Mean (SD)	Min, Max	Mean (SD)	Min, Max
Mean 24‐h urine volume	1995.6 (780.4)	700, 3880	2089.0 (791.7)	730, 3780	1831.9 (742.8)	700, 3880
Male	2083.3 (788.5)	825, 3880	2077.7 (797.7)	825, 3600	2092.1 (793.5)	1070, 3880
Female	1890.2 (766.0)	700, 3780	2101.5 (798.5)	730, 3780	1467.7 (486.0)	700, 2610
Mean 24‐h Cr exc (mg/d)
Measured 24‐h	1510.9 (534.7)	303, 2925	1682.1 (495.4)	704, 2925	1211.3 (469.7)a, b	303, 2481
Estimated 24‐h (algorithm)	1689.0 (451.5)	830, 2652	1685.4 (470.7)	830, 2610	1695.3 (422.2)	997, 2652
Male
Measured 24‐h	1809.4 (461.8)	867, 2925	2017.4 (380.4)	1366, 2925	1482.5 (387.3)	867, 2481
Estimated 24‐h	1993.9 (286.9)	1496, 2652	2026.9 (272.3)	1606, 2611	1942.1 (308.2)	1496, 2652
Female
Measured 24‐h	1152.7 (373.3)	303, 2175	1313.3 (310.4)	704, 2175	831.5 (269.3)	303, 1311
Estimated 24‐h	1323.2 (322.3)	830, 2231	1309.9 (336.5)	830, 2231	1349.7 (301.2)	997, 1974
Mean 24‐h Na exc (mEq/d)
Measured 24‐h uncorrected	143.3 (71.9)	22, 349	156.2 (73.8)	28, 349	120.6 (63.3)a	22, 307
Measured 24‐h corrected	162.3 (75.8)	29, 357	157.2 (76.1)	29, 357	171.2 (75.6)b	40, 328
Male
Measured 24‐h uncorrected	157.3 (71.9)	28, 349	172.0 (74.7)	28, 349	134.2 (61.9)	33,307
Measured 24‐h corrected	175.1 (73.2)	29, 337	173.7 (75.0)	29, 337	177.2 (72.0)	40, 328
Female
Measured 24‐h uncorrected	126.5 (69.0)	22, 306	138.9 (69.8)	48, 306	101.7 (62.2)	22, 268
Measured 24‐h corrected	146.9 (76.9)	52, 357	139.0 (74.2)	52, 357	162.7 (74.2)	72, 324

^aComparing complete vs incomplete collectors, P < .0001 for measured 24‐h Cr exc; P = .02 for measured 24‐h uncorrected Na exc, by independent samples t test.

^bComparing within complete collectors and within incomplete collectors, P < .0001 difference between measured and estimated 24‐h Cr exc and for difference between uncorrected and corrected 24‐h Na exc, by paired samples t test.

Among complete collectors, mean estimated and measured 24‐hour creatinine values were very similar to each other (1685 and 1682 mg/d, respectively, P = .89), as were the uncorrected and corrected 24‐hour sodium excretion values (156 and 157 mEq/d, respectively, P = .70). In contrast, among incomplete collectors, the mean measured 24‐hour creatinine excretion was significantly lower than the estimated mean, 1211 vs 1695 mg/d, respectively (P < .0001), and, similarly, the uncorrected 24‐hour sodium excretion was significantly lower than the corrected sodium excretion, 121 vs 171 mEq/d, respectively (P < .0001).

Participants were then stratified into subgroups based on whether they were or were not taking a diuretic at the time of the urine collections. (Table 3). As was observed for the entire cohort, the measured 24‐hour creatinine and sodium excretion values were lower among incomplete collectors than among complete collectors in both diuretic users and non‐users. Among complete collectors, the measured and corrected 24‐hour excretion values were very similar to each other, both in the diuretic‐ and non‐diuretic‐treated subgroups. In contrast, among incomplete collectors, the measured 24‐hour excretion values were lower than in complete collectors in both diuretic users and non‐users. After correction, the corrected values for 24‐hour sodium excretion similarly approximated the values of complete collectors regardless of diuretic status.

Table 3.

Estimated and measured mean 24‐h urinary creatinine and sodium excretion by diuretic use

	Entire sample		Complete collectors		Incomplete collectors		P b
	Mean (SD)	P a	Mean (SD)	P a	Mean (SD)	P a
Mean 24‐h Cr exc (mg/d)	N = 99		N = 63		N = 36
Measured 24‐h	1510.9 (534.7)	<.0001	1682.1 (495.4)	.89	1211.3 (469.7)	<.0001	<.0001
Estimated 24‐h (algorithm)	1689.0 (451.5)		1685.4 (470.7)		1695.3 (422.2)		.92
On Diuretic	N = 60		N = 33		N = 27
Measured 24‐h	1499.9 (480.7)	<.0001	1664.5 (449.0)	.89	1298.6 (446.9)	<.0001	.003
Estimated 24‐h	1701.3 (428.4)		1660.2 (425.1)		1751.4 (435.0)		.42
Not on Diuretic	N = 39		N = 30		N = 9
Measured 24‐h	1527.9 (614.9)	.008	1701.5 (549.1)	.75	949.1 (461.5)	<.0001	.001
Estimated 24‐h	1670.2 (490.1)		1713.2 (522.2)		1526.8 (349.9)		.32
Mean 24‐h Na exc (mEq/d)	N = 99		N = 63		N = 36
Measured 24‐h uncorrected	143.3 (71.9)	<.0001	156.2 (73.8)	.70	120.6 (63.3)	<.0001	.02
Measured 24‐h corrected	162.3 (75.8)		157.2 (76.1)		171.2 (75.6)		.38
On Diuretic	N = 60		N = 33		N = 27
Measured 24‐h uncorrected	145.5 (65.3)	<.0001	156.5 (64.7)	.88	132.2 (64.6)	<.0001	.15
Measured 24‐h corrected	167.4 (71.9)		156.0 (64.0)		181.3 (79.5)		.18
Not on Diuretic	N = 39		N = 30		N = 9
Measured 24‐h uncorrected	139.9 (81.8)	.01	156.0 (83.7)	.54	86.0 (46.1)	.002	.02
Measured 24‐h corrected	154.4 (81.8)		158.4 (88.6)		140.9 (55.5)		.58

^aComparing within entire sample, complete collectors and within incomplete collectors, difference between measured and estimated 24‐h Cr exc and for difference between uncorrected and corrected 24‐h Na exc, by paired samples t test.

^bComparing complete vs incomplete collectors by independent samples t test.

The individual values for 24‐hour creatinine excretion are shown in the scatterplot in Figure 1. Low values, notably below 800 mg/d, are observed among measured values, but not among estimated values.

Figure 1.

Relationship between estimated 24‐h creatinine excretion and measured 24‐h creatinine excretion in (A) incomplete collectors and (B) complete collectors

The individual values for 24‐hour sodium excretion are displayed in Figure 2. Among incomplete collectors, 11 of 36 participants had a measured sodium excretion in the desirable range of <100 mEq/d. However, the corrected value was <100 mEq in only 7.

Figure 2.

Relationship between measured corrected 24‐h sodium excretion and measured uncorrected 24‐h sodium excretion in (A) incomplete collectors and (B) complete collectors

For 24‐hour creatinine excretion in incomplete collectors (Figure 3A), the mean difference between measured and estimated values was great (−484.045, CI −559.440, −408.650), with the measured values substantially lower than those estimated; the width of the 95% limits of agreement was −920.794 and −47.296. In contrast, for complete collectors (Figure 3B), the mean difference between measured and estimated creatinine values was −3.310 (CI: −50.566, 43.946). The 95% limits of agreement around the mean difference −3.310 was −371.078 and 364.457.

Figure 3.

Bland‐Altman plots showing the difference between measured 24‐h creatinine excretion and estimated 24‐h creatinine excretion in (A) complete collectors and (B) incomplete collectors. This difference for each individual is plotted against the mean of the estimated and measured 24‐h creatinine excretion

For 24‐hour sodium excretion in incomplete collectors (Figure 4A), the mean difference between measured uncorrected and measured corrected values was −50.550 (−62.088 and −39.013), with the measured uncorrected values lower than those corrected; the width of the 95% limits of agreement around −50.550 was −117.386 and 16.285. In contrast, for complete collectors (Figure 4B), the mean difference between measured uncorrected and measured corrected sodium values was −0.907 (CI: −5.661, 3.848). The 95% limits of agreement around the mean difference of −0.907 was −37.910, 36.097.

Figure 4.

Bland‐Altman plots showing the difference between measured uncorrected 24‐h sodium excretion and measured corrected 24‐h sodium excretion in (A) complete collectors and (B) incomplete collectors. This difference for each individual is plotted against the mean of the measured uncorrected and measured corrected 24‐h sodium excretion

4. DISCUSSION

Incomplete 24‐hour urine collections are a widespread problem in measuring excretion of sodium or any other analyte. In this study, participants whose 24‐hour urine collection was determined to be incomplete based on a lower than predicted 24‐hour creatinine excretion appeared misleadingly to have a lower 24‐hour sodium and creatinine excretion than those who provided a complete collection. There is clearly no reason to believe that incomplete collectors truly consume less sodium and excrete less creatinine. This artifactual difference was eliminated when the 24‐hour creatinine and sodium excretion were corrected for the degree of incompleteness of collection, as remedied by the ratio of estimated to measured creatinine excretion. Thus, for example, in a male who weighs 150 pounds whose 24‐hour urine contains 75 mEq of sodium, if the creatinine excretion was 850 mg rather than the estimated value of 1700 mg/d, the corrected sodium excretion would be 150 mEq rather than the 75 mEq contained in the incomplete collection.

The accuracy of values predicted from the spot urine was evident both in participants who were not taking a diuretic and, importantly, in participants who were taking a thiazide and/or potassium‐sparing diuretic. We did not assess the accuracy of the equation in participants taking a loop diuretic.

In this study, 36% of the participants whose measured sodium excretion was <100 mEq/d were in fact excreting >100 mEq. This finding disturbingly suggests that in many individuals, the finding of low 24‐hour sodium excretion may be artifactual due to incompleteness of the urine collection. Greater accuracy can be achieved by the simple arithmetic adjustment of the results by the ratio of estimated to measured creatinine excretion.

Without correcting for the low creatinine content indicative of incomplete collection, the measured excretion of sodium or any other analyte underestimates the actual excretion, with potentially important clinical implications. Given the high frequency of incomplete collections, routine use of this adjustment, that requires knowledge only of age, weight, sex, and race, could play an important role in better interpreting results of 24‐hour urine collections in both clinical and research realms.

Results of previous studies support our findings. For example, in a study reported by James et al,14 among 144 participants who submitted a 24‐hour urine collection annually for 9 years, the mean 24‐hour creatinine excretion was 1642 mg in men (average weight 78 kg) and 1041 mg in women (average weight 66 kg). These means were somewhat similar to the mean measured 24‐hour creatinine excretions in the current study likely reflecting a mixture of complete and incomplete collections that similarly understated the true mean 24‐hour creatinine excretion. In the current study, for the entire cohort including both complete and incomplete collectors, the 24‐hour creatinine excretion, 1809 mg among men and 1153 mg among women, appeared similar to that reported by James. The mean creatinine excretion is actually higher after adjusting for the effect of incomplete collection by either correcting the 24‐hour measured creatinine values (resulting in a mean of 1994 mg among men and 1323 mg among women) or looking only at complete collectors (2017 mg among men and 1313 mg among women).

Previous studies have also constructed equations to estimate 24‐hour creatinine excretion. Those equations were constructed from the variables of height, weight, age, and sex.15, 16, 17 Development of our equation differed from those by Tanaka and by Kawasaki in several ways. Although similarly based on weight, age, and sex, our equation was not based on height. Also, it was constructed from an American rather than a Japanese sample and included the additional independent variable of race.

It also differed in the manner in which incomplete 24‐hour collections were excluded from the data used to construct the equation. Clearly, inclusion of incomplete collections would result in an equation that underestimates 24‐hour creatinine excretion. In the Tanaka study, assessment of completeness of collection was based on interview.16 In the Kawasaki study, it was based on the result of collections from 3‐5 days, with no mention of handling of incomplete collections.17 The INTERSALT study also likely included many undercollectors as their criterion for inclusion was a 24‐hour volume exceeding a mere 250 cc.15 As a consequence, 24‐hour creatinine might have been underestimated in all three studies.

A study by Ix et al18 used pooled data from three previous studies to develop creatinine estimating equations, followed by validation of the equation using data from three other studies. Although this study had a large sample size for developing and validating the equation, the study participants included in both the development and validation datasets differed from the participants in the current study. Their pooled sample included participants from clinical trials, most of whom had kidney disease. Further, the completeness of the 24‐hour collections was not the main purpose of those studies and there was little attempt to identify incomplete collections. They excluded only samples whose estimated creatinine excretion was either extremely low or extremely high and biologically implausible, that is measured creatinine excretion rate <350 or >3500 mg/d.18

Nonetheless, comparing the performance of the two new equations by Ix et al to four previously published equations,19, 20, 21, 22 demonstrated, with the exception of the Cockcroft‐Gault equation, generally good agreement. However, again, the participants in the four previous studies included many individuals with chronic kidney disease.21, 22

In our study, determination of incomplete collection was based on previously recommended cutoff values of 24‐hour creatinine content <20 mg/kg in men and <15 mg/kg in women.6 Further, the accuracy of the equation was also assessed by observing the strength of its correlation with muscle mass, measured by bioelectrical impedance, which served as a gold standard that is independent of the vagaries of the 24‐hour urine collection.23 The strong relation between urine creatinine excretion and muscle mass has previously been shown.10 In our previous study, the correlation of the equation‐derived estimate of 24‐hour creatinine excretion with muscle mass was found to be stronger than that observed with the actual 24‐hour collections, clearly due to correction for undercollection.3

4.1. Limitations

Body weight is one of the variables used in predicting creatinine excretion. Clearly, different individuals of the same weight can differ in their proportion of adipose and muscle mass. In obese patients, the equation might overestimate, and in very muscular patients, the equation might underestimate, the actual creatinine excretion. This confounding factor is partly mitigated by the increase in muscle mass that typically accompanies an increase in fat mass; for example, a 30 kg increase in fat mass is accompanied by a 12 kg increase in muscle mass.24 Despite this limitation, use of the estimated creatinine excretion considerably reduces the inaccuracy resulting from incomplete collections.

The results of this study cannot be extrapolated to individuals with advanced renal disease or to those whose sodium and food intake or metabolic or hemodynamic functions are not in a steady state. Finally, our study population consisted of white and African‐American individuals. The generalizability of the data to other populations remains to be assessed.

5. CONCLUSION

Undercollection is a frequent problem in using the 24‐hour urine collection as the gold standard for determining the 24‐hour excretion of sodium or any other analyte. In undercollectors, the measured 24‐hour excretion of any analyte substantially understates the true 24‐hour excretion.

In practice, one often cannot differentiate an individual who submits a complete collection from one who submits an incomplete one. We have demonstrated that in measuring the 24‐hour excretion of sodium, or presumably of any analyte, correction for incompleteness of collection by adjusting for the ratio of estimated to measured creatinine excretion provides a simple and important modification that compensates for undercollection and can considerably improve interpretation of results of the standard 24‐hour collection. Therefore, adjustment for the estimated creatinine excretion merits consideration for routine clinical use in reporting the results of 24‐hour urine collections of sodium or any other analyte being measured.

CONFLICT OF INTEREST

The authors are recipients of a patent related to determining excretion of sodium and other analytes. The authors declare no competing financial interests.

AUTHOR CONTRIBUTIONS

SJM and LMG: conceived and designed the study and contributed to interpretation of data. SJM: was involved in the acquisition of data. LMG: analyzed the data. Both authors contributed to drafting the manuscript and to critically reviewing and revising the manuscript. Both authors approved the final version of this submission.

Mann SJ, Gerber LM. Addressing the problem of inaccuracy of measured 24‐hour urine collections due to incomplete collection. J Clin Hypertens. 2019;21:1626–1634. 10.1111/jch.13696