The 24-hour pad test in continent women and men: normal values and cyclical alterations

BJOG: an International Journal of Obstetrics and Gynaecology June 2003, Vol. 110, pp. 567 – 571

The 24-hour pad test in continent women and men: normal values and cyclical alterations E. Karantanis, R. O’Sullivan, K.H. Moore* Objective To obtain control values for the 24-hour pad test in a wide age range of continent women using accurate weighing scales and to compare the results obtained from pantyliners and pads in women and men. Design Prospective anonymous study. Setting A Sydney Tertiary Urogynaecology Unit. Population One hundred and forty continent women and 14 continent men. Methods A 24-hour pad test was performed in 120 continent women of widely varying ages, in whom hormonal status and exercise habits were documented. These continent women wore a standardised pantyliner for 24 hours and a high precision beam balance (accuracy 0.1 g) was used to measure the loss on the pad. To assess any variation in pad weights with differing pads, 20 female volunteers undertook the 24hour test firstly with pantyliners then larger pads. To understand evaporative qualities of the two types of pads, one male wore each type of pad, instilled with 5 mL normal saline for eight hours overnight. Furthermore, to understand the contribution of vaginal secretions, a group of male volunteers performed a 24-hour test with pantyliners followed by pads. Main outcome measures Pad weight, with regard to hormonal status, exercise, pad type and gender. Results The median age of subjects was 48 (interquartile range [IQR] 32 –60), with a median pad weight gain of 0.3 g (IQR 0.2 – 0.6; 95th centile 1.3 g). Subgroup analysis showed no significant trends for pad loss in relation to menopause status, use of hormone replacement therapy or hormonal contraception and exercise status during the 24-hour period. Control values for pantyliners were not significantly different from those for continence pads. In addition, normal values in 14 males showed similar results, regardless of pad type. Conclusions The response rate of 39% might limit the applicability of the results. However, our finding that women lose only 0.3 g of vaginal secretions in 24 hours is much lower than previously reported. This might arise from the use of a highly accurate beam balance and the recruitment of a large sample of women with widely varying ages. This result might lower the threshold for objective diagnosis of urinary incontinence and alters the pad test definition of ‘cure’. INTRODUCTION Both the 1 and 24-hour pad tests have been employed as outcome measures in the field of urinary incontinence for several years. As the sensitivity and reproducibility of the 1 hour pad test has been increasingly called into question1 – 3, the 24-hour pad test will perhaps become more widely used. When adopting a new outcome measure, ‘reference’ values should be clearly established. Three studies have revealed that median vaginal fluid loss in continent women may be as high as 2.6– 7.0 g, with an upper confidence limit (CL) of 5.5 – 8.0 g in 24 hours4,5 or 14 g over 48 hours6. The majority of women in these studies were young physiotherapy or nursing students and may have been ovulating4,6. We question therefore whether younger

Pelvic Floor Unit, St George Hospital, University of New South Wales, Sydney, Australia * Correspondence: Assoc Prof K. H. Moore, Level 1, Clinical Sciences Building, St George Hospital, Kogarah, New South Wales 2217, Australia. D RCOG 2003 BJOG: an International Journal of Obstetrics and Gynaecology doi:10.1016/S1470-0328(03)02115-3

women with plentiful cervical mucus are acceptable controls for incontinent women who are usually older and not ovulating. Assuming evaporative loss of moisture is not significantly greater on underwear than on pads, we find such results hard to reconcile with our ordinary clinical experience, as we suspect that 3 –7 mL of fluid would dampen underwear and evoke some symptoms. Finally, while vaginal secretions are a normal part of female physiology, few studies have attempted to characterise the amount of loss in relation to variables such as menopause status, hormone use and exercise patterns. Therefore, the primary aim of this study was to conduct 24-hour pad tests to obtain reference values based on a larger sample than previously reported (at least 100 continent women) and to recruit women from a wide range of ages. In particular, we aimed to enrol a significant proportion of postmenopausal continent women to approximate the median age of incontinent subjects. Furthermore, we used a high precision beam balance, accurate to 0.1 g, rather than the spring balance used by previous authors4,6. One study that obtained normal values for postmenopausal women gave little description of the weighing scales used7. www.bjog-elsevier.com

568

E. KARANTANIS ET AL.

The age range in this study was 45 – 58 years, which we considered to be relatively narrow considering the wide age range of women who suffer from incontinence. All pad weighing in our study was performed by trained staff rather than by patients themselves4,6. Our secondary aim was to determine the relationship between pad weight gain and hormonal status (i.e. the use of hormone replacement therapy in postmenopausal women, the use of hormonal contraception and the timing of the menstrual cycle in premenopausal women and the performance of exercise in the total sample). During the analysis of the data, we realised our normal values were considerably lower than that reported by any previous author. As will be described in the Methods, we had used a small pantyliner to facilitate anonymous postal returns, whereas previous authors had not specified the exact pad type used4 – 7. Therefore, we felt obliged to measure any difference between ordinary continence pads and pantyliners in continent women, which we undertook in Experiment 2. A final experimental dilemma was the lack of information regarding the role of perspiration, vaginal secretions and the rate of evaporation in producing the final pad weight. To clarify these problems, we enrolled a group of men in our study to eliminate the effect of vaginal secretions upon pad weight. Finally, we persuaded these men to wear both a pantyliner and a sanitary pad for 24 hours.

METHODS For Experiment 1 (general controls/pantyliner), recruitment of continent women was carried out by approaching the following groups including (1) attendees of a menopause seminar, (2) senior midwives and nursing staff, (3) clerical staff, (4) women from an osteoporosis aerobics group, (5) hospital volunteers, (6) nurse continence advisors and (7) medical students. The study purpose was discussed: it was stressed that only continent volunteers were required. In endeavouring to provide strict confidentiality and anonymity, no attempt was made to question subjects about their continence, but it was emphasised that those with incontinence should not return the pad weighing test. The study took place in the Australian winter, with average temperatures reaching an approximate maximum of 16jC (61jF), so that the component of perineal perspiration would not be atypically large, compared with other countries. Each subject was given a stamped self-addressed test parcel containing three pads in a transparent snaplock bag, information sheets and a questionnaire. The pads used were ‘Johnson & Johnson Carefree Pantyliners’. These were chosen in preference to thicker pads because it was felt that the less bulky pads were easier to send in a mail parcel, more comfortable and thus more likely to optimise the compliance rate. Pads were purchased in 45 pad boxes, of

which two pads and two adhesive strips were randomly sampled per box to ensure weights were identical. Each pad weighed 2.4 g without adhesive and was approximately 50 mm wide, 150 mm long and 3 mm thick and contained no superabsorber gel. Three pads were then placed in the snaplock bag and weighed with a beam balance accurate to 0.1 g (ADEK Melbourne, Australia, Model 1200A). The pre-test weight was noted and labelled on the bag. The parcels were completely anonymous and unmarked but were distributed sequentially to the different groups described earlier (Groups 1 –7) with several days separating each group. In this way, parcels tended to return in separate ‘waves’ so that approximate response rates could be tracked for the different groups. Women were instructed to wear between one and three of the supplied pads in a 24-hour period and ensure that all three pads and their adhesives were sealed in the snaplock bag whether used or unused. They were required to abandon the study if intercourse occurred. Women were asked to complete a questionnaire and return the test parcel by mail as soon as possible after performing the 24-hour test. The questionnaire ascertained whether the patient leaked urine, her age, menopause status, and whether hormone replacement or hormonal contraception were being used. In addition, women were asked to indicate if they had a prior hysterectomy with or without bilateral oophorectomy or if they were using vaginal creams of any sort. Premenopausal women who menstruated were asked the duration since the first day of their menstrual period. For the sake of simplicity, questions about cycle length and regularity were not included. To measure the effect of exercise upon pad weight, women in the initial phase of the study were asked to avoid exercise for the 24 hours of the pad test (called Group 1). After distributing 50% of the test parcels, the instruction sheet/questionnaire was modified to allow exercise (Group 2). Women were asked to undertake all normal daily activities and to indicate on the questionnaire whether they had exercised. Exercise was loosely defined as any recreational activity that caused sweating to occur or increased the heart rate. By this stage of Experiment 1, the 39% return rate was apparent (see Results). Therefore, to achieve >100 completed tests, we anticipated needing to distribute >300 test parcels. The snaplock bags in the test parcel were weighed in the unit within three days of their receipt, in the knowledge that pads stored in sealed bags maintain their weight for at least seven days6. Bags were checked for the number of pads and adhesives, weighed and the net weight gain calculated. If an adhesive strip or pad was missing, this was adjusted for in the calculation. The trend between pad weight and the timing of the menstrual cycle was analysed by comparing the pad weights of women whose last menstrual period occurred 10 – 18 days prior to the test (periovulatory period), with those outside of this range. For Experiment 2, we tested whether our normal values obtained by pantyliner measurements differed significantly D RCOG 2003 Br J Obstet Gynaecol 110, pp. 567 – 571

THE 24-HOUR PAD TEST IN CONTINENT WOMEN AND MEN

if we used ordinary pads (of greater surface area, thickness and absorbancy). Thus, a smaller sample of women were asked to wear the pantyliner according to the current method for 24 hours, then wear the thicker larger ‘Tena Lady Normal’ pads. These were 7 mm thick, 260 mm long and 110 mm wide and weighed 14.5 g without adhesives or wrapping. They contained absorber gel. The majority of pads were also collected during the winter months. Early results for the first five women in Experiment 2 (pad vs pantyliner) suggested a probable mild increase in pad weight when using thicker pads. This prompted a third experiment (male absorption test) to explore the hypothesis that thicker pads have a slower evaporation rate. One male volunteer wore a pantyliner for eight hours overnight, then an ordinary pad the following night. Nighttime was chosen to minimise any potential contribution of sweat to the weight of the pad. Each night, the relevant pad was instilled with 5 mL normal saline, removed the next morning and reweighed. In Experiment 4, the pantyliner was compared to pad in males. Over the subsequent winter months, male volunteers were approached to wear a pantyliner for 24 hours followed by a pad. The purpose was firstly to study the relative contribution of perspiration to pad weight by eliminating female vaginal secretions from the experiment. We also compared the relative pad weights of pantyliner and pad in men. In all four experiments, the methodology and statistical analysis was as for Experiment 1. Data were entered on an Excel database and analysed with Arcus Quickstat statistical software package. As data were not normally distributed, they are presented as median and interquartile range (IQR). Logarithmic transformation did not normalise the distribution, rendering the calculation of 95% upper CL inappropriate. As results closer to zero were likely to be biologically most accurate, the data were ranked in ascending order and the upper 5% of data were removed as

569

statistical outliers. The highest remaining figure was thus represented as the 95th centile. Group differences were compared by the Mann – Whitney U test. Comparison of the different pad types within groups was calculated by Wilcoxon’s signed rank test.

RESULTS In Experiment 1, Random sampling of the two pantyliners from each of eight boxes revealed a variation of no more than 0.1 g. The 320 parcels were distributed as follows: menopause seminar (80), senior midwives and nursing staff (65), clerical staff (40), an osteoporosis aerobics group (35), hospital volunteers (30), nurse continence advisors (15), medical students (35) and other (20). A total of 124 women returned their parcels, giving a 39% response rate, although due to anonymity the response rate of the various groups could not be ascertained exactly. Each of the cluster groups appeared to have similar rates of return, although this could not be quantified. Four parcels were discarded, one because the women indicated they leaked urine, one because of an open bag and two because they used vaginal creams. The median age for the 120 participants was 48 years (IQR 32 – 60, range 20– 80). The median pad test weight gain was 0.3 g (IQR 0.2– 0.7, 95% upper CL 0.4 g). Table 1 shows that 60% of participants (72/120) were Group 1 (avoided exercise). The pad test differences between Groups 1 and 2 were not significant, the same being true for the subset analysis within Group 2 (those who did or did not chose to exercise). Therefore, subsequent analysis of hormonal status was performed upon pooled data from both groups. The pad test loss did not differ significantly between preand postmenopausal women. In postmenopausal women,

Table 1. Subgroup comparisons of 24-hour pad weight gains. Values given are n, (subtotal of n), median [IQR], 95th centile and P value. Subgroup variable

n

Pad weight

95th centile (g)

Group 1 (no exercise)

72

0.3 [0.2 – 0.6]

1.3

Group 2 (exercise optional) Did exercise Did not exercise

48 (21) (27)

0.25 [0.1 – 0.5] 0.3 [0.1 – 0.4] 0.2 [0.1 – 0.3]

1.2

Hormonal effects Postmenopause Hormone replacement therapy No hormone replacement therapy Premenopause Hormonal contraception No hormonal contraception Cycle day 10 – 18# Non-ovulatory phase

57 (23) (34) 63** (14) (46) (25) (19)

0.3 0.3 0.25 0.3 0.3 0.3 0.2 0.5

1.3

[0.2 – 0.5] [0.2 – 0.6] [0.1 – 0.5] [0.2 – 0.7] [0.1 – 1.1] [0.2 – 0.7] [0.1 – 0.4] [0.3 – 1.2]

0.08 0.8

1.3

* P value for total pre- vs postmenopause. ** Three premenopausal women did not complete the questionnaire — they left the question about contraception blank. # Missing data in two subjects.

D RCOG 2003 Br J Obstet Gynaecol 110, pp. 567 – 571

P value, Mann – Whitney U test

0.3 0.3* 0.8 0.008

570

E. KARANTANIS ET AL.

Table 2. 24-hour pad weight comparisons between continent women and men. Values given are median [IQR], 95th centile and P value. Pantyliner (g)

Pad (g)

P value, Wilcoxon’s signed ranks test

Women (n == 21) Weight 95th centile

0.4 [0.2 – 0.6] 1.0

0.5 [0.4 – 1.0] 1.2

0.12

Men (n == 14) Weight 95th centile

0.25 [0 – 0.25] 0.6

0.25 [0.1 – 0.4] 0.5

0.82

0.1

0.006

P value, Mann – Whitney U test

there was no significant difference in the pad weight of hormone replacement therapy users (Table 1). Similarly, there was no significant difference in pad test weight between those who did and did not use hormonal contraception. However, in premenopausal women not using hormonal contraception, the pad weights were significantly lower in periovulatory women (Table 1). In Experiment 2, the initial five results in a new group of volunteers, showing a difference between pantyliner and pad weights (discussed in Methods), were not representative of the eventual total sample. Therefore, all results for this experiment were pooled (Table 2). Of 22 women, one postmenopausal woman was excluded because she indicated she leaked urine. Median pantyliner weights were 0.4 g and the pad weight median was 0.5 g (no significant difference, Table 2). In Experiment 3, the overnight saline instillation could only be conducted on one male volunteer. After 5 mL of saline instillation followed by overnight exposure to body temperature in the genital region, the pantyliner weight decreased by 2.5 g. In comparison, the pad weight decreased by 0.5 g. In Experiment 4, of 14 men who wore pantyliners followed by pads, the median weight gain was identical for each type of protection (Table 2). However, pad weight gain was significantly greater for women than men (0.5 vs 0.25 g, P ¼ 0.006, Mann – Whitney U test) (Table 2), suggesting that vaginal secretions accounted for about half of the pad loss in continent subjects.

DISCUSSION Our aim was to clarify the precise range of values for the 24-hour pad test results in older continent women. Although previous studies have provided control data, sample sizes were generally smaller and measurements were made on simple weighing scales. For example, Lose et al.5 recruited 23 continent women (age 34 – 69 years) and found a median pad test result of 4 g (99% upper CL 8 g).

The accuracy of the scale used was 1 g (personal communication). Versi et al.6 and Mouritsen et al.4 tested 24 ‘young nursing and physiotherapy students’ (mean age not stated) and 25 nursing and physiotherapy staff (median age 41), respectively. They observed a mean pad gain of 7 (95% upper CL 14 g) and 2.6 g (range 0 –7 g, 95% upper CL 5.5 g) in 48 hours, respectively. In the first study, spring balances were used, the accuracy of which was not mentioned, while in the latter study, women weighed their own pads at home using a simple balance. More recently, Ryhammer et al.7 conducted 24-hour pad tests on 78 continent postmenopausal women but the age range was quite narrow (45 – 57, mean 50 years). They reported a mean pad test result of 3.1 g (range 0– 9 g). The scale used in this study was accurate to only 1 g and the pad type was unspecified. In the present study, we have undertaken 120 pad tests in women who had no symptoms of incontinence and pad weight gains were substantially less than that observed in three other studies. However, the sample sizes of other studies were smaller and most used spring balance scales of unknown accuracy. Obviously, if the pre-test pad is weighed inaccurately and the loss after 24 hours is also weighed inaccurately, then the compound error for two to three pads could easily total 3 g. Our response rate of 39% was poor but many volunteers expressed distaste at the prospect of sending pads containing vaginal secretions back to be examined and weighed. Trends for the seven different groups did not suggest any discernable difference in response rates for the different subsets of women. We also investigated whether exercise, menopause, hormone replacement therapy, hormonal contraception or the menstrual cycle was associated with variations in pad weight gain. The finding that no differences were found between pre- and postmenopausal results challenges the premise that vaginal secretions are greater in premenopausal women. Subgroup analysis revealed no significant effect of exercise upon pad weight gain. Our results are in keeping with the previous finding of Versi et al.6 in physiotherapists. The finding that periovulatory pad weight gain was less than the remainder of the menstrual cycle (excepting menstruation) is difficult to interpret. One might speculate that the quality of ovulatory cervical mucus (spinnbarkeit) renders it less likely to actually run on the pad. The studies performed in men have shown that the contribution from vaginal secretions is about half of the total pad weight for controls, the remainder probably due to peri-anal fluid or perspiration. In summary, the previously reported upper limit of normal for the 24-hour pad test in continent women of 5.5 g is not supported by our study. Instead, our finding gives a median pad weight gain for continent women of just 0.3 g (95% upper CL 0.4 g) and a median of 0.25 g for continent men. D RCOG 2003 Br J Obstet Gynaecol 110, pp. 567 – 571

THE 24-HOUR PAD TEST IN CONTINENT WOMEN AND MEN

We chose pantyliners for this study for less perineal discomfort and hopefully better compliance and to facilitate ordinary postal return and anonymity. However, the response rate of 39% was disappointing and could limit the overall applicability of the findings. We proceeded to check whether pad loss varied with the type of pad used, as this could explain the variation between our findings and those of other authors. The pad weight gains for the thinner, lighter ‘Johnson & Johnson Pantyliners’ and the larger, heavier ‘Tena Lady Normal’ pads did not differ significantly from the change in pad weight seen when using the ‘Tena Lady Normal’ pads. In most previous studies, a pack of five pads was given to each test subject. We chose only three pads per test to increase compliance. Even if we had used five pads, our control value over 24 hours would not likely exceed 0.4 g and would have made the study more burdensome for these continent women. In conclusion, we recommend that the results of the 24hour pad test should always be measured on accurate weighing scales, particularly in the research setting, but the type of pantyliners or pads used are immaterial so long as they are clearly specified. The data from this study suggest a new value for continent controls and that the definition of ‘objective cure’ of incontinence might need to be revised. If our lower cutoff limit of ‘mild incontinence’ is accurate (taking into account the low response rate), the

D RCOG 2003 Br J Obstet Gynaecol 110, pp. 567 – 571

571

future attempts to define mild moderate and severe categories for the 24-hour test should be undertaken with this lower limit in mind.

References 1. Simons AM, Yoong WC, Buckland S, Moore KH. Inadequate repeatability of the one-hour pad test: the need for a new incontinence outcome measure. Br J Obstet Gynaecol 2001;108(3):315 – 319. 2. Fantl JA, Harkins SW, Wyman JF, Choi SC, Taylor JR. Fluid loss quantitation test in women with urinary incontinence: a test – retest analysis. Obstet Gynecol 1987;70(5):739 – 743. 3. Schick E, Joilvet-Tremblay M. Detection and quantification of urine of urine loss: the pad-weighing test. In: Corcos J, Schick E, editors. The Urinary Sphincter. Montreal: Marcel Dekker, 2001:275 – 283. 4. Mouritsen L, Berild G, Hertz J. Comparison of different methods for quantification of urinary leakage in incontinent women. Neurourol Urodyn 1989;8:579 – 587. 5. Lose G, Jorgensen L, Thunedborg P. 24-hour home pad weighing test versus 1-hour ward test in the assessment of mild stress incontinence. Acta Obstet Gynecol Scand 1989;68(3):211 – 215. 6. Versi E, Orrego G, Hardy E, Seddon G, Smith P, Anand D. Evaluation of the home pad test in the investigation of female urinary incontinence. Br J Obstet Gynaecol 1996;103(2):162 – 167. 7. Ryhammer AM, Laurberg S, Djurhuus JC, Hermann AP. No relationship between subjective assessment of urinary incontinence and pad test weight gain in a random population sample of menopausal women. J Urol 1998;159(3):800 – 803. Accepted 4 March 2003