Do Stool Form and Frequency Correlate With Whole ...

3 downloads 0 Views 301KB Size Report
Such testing allows. Do Stool Form and Frequency Correlate With Whole-Gut and Colonic Transit ? Results From a Multicenter Study in Constipated Individuals ...
Do Stool Form and Frequency Correlate With Whole-Gut and Colonic Transit? Results From a Multicenter Study in Constipated Individuals and Healthy Controls Richard J. Saad, MD1, Satish S.C. Rao, MD2, Kenneth L. Koch, MD3, Braden Kuo, MD4, Henry P. Parkman, MD5, Richard W. McCallum, MD6, Michael D. Sitrin, MD7, Gregory E. Wilding, PhD7, Jack R. Semler, PhD8 and William D. Chey, MD1

OBJECTIVES:

Despite a lack of supportive data, stool form and stool frequency are often used as clinical surrogates for gut transit in constipated patients. The aim of this study was to assess the correlation between stool characteristics (form and frequency) and gut transit in constipated and healthy adults.

METHODS:

A post hoc analysis was performed on 110 subjects (46 chronic constipation) from nine US sites recording stool form (Bristol Stool Scale) and frequency during simultaneous assessment of whole-gut and colonic transit by wireless motility capsule (WMC) and radio-opaque marker (ROM) tests. Stool form and frequency were correlated with transit times using Spearman’s rank correlation. Accuracy of stool form in predicting delayed transit was assessed by receiver operating characteristic analysis.

RESULTS:

In the constipated adults (42 females, 4 males), moderate correlations were found between stool form and whole-gut transit measured by WMC (r = − 0.61, P < 0.0001) or ROM ( − 0.45, P = 0.0016), as well as colonic transit measured by WMC ( − 0.62, P < 0.0001). A Bristol stool form value < 3 predicted delayed whole-gut transit with a sensitivity of 85% and specificity of 82% and delayed colonic transit with a sensitivity of 82% and specificity of 83%. No correlation between stool form and measured transit was found in healthy adults, regardless of gender. No correlation was found between stool frequency and measured transit in constipated or healthy adults. The correlation between stool frequency and measured transit remained poor in constipated adults with < 3 bowel movements per week.

CONCLUSIONS: Stool form predicts delayed vs. normal transit in adults. However, only a moderate correlation

exists between stool form and measured whole-gut or colonic transit time in constipated adults. In contrast, stool frequency is a poor surrogate for transit, even in those with reduced stool frequency. Am J Gastroenterol advance online publication, 3 November 2009; doi:10.1038/ajg.2009.612

INTRODUCTION Constipation is a prevalent, symptom-based disorder (1,2). The symptoms commonly reported by constipated patients include infrequent bowel movements (less then three bowel movements a week), the passage of hard or lumpy stool, straining with bowel movements, a sensation of incomplete evacu-

ation following a bowel movement, a sensation of anorectal blockage or obstruction during bowel movements, and use of manual maneuvers to facilitate the passage of a bowel movement. Specialized testing can be conducted to better define the underlying physiological or structural abnormalities that are responsible for such symptoms (3–5). Such testing allows

1

Department of Internal Medicine, University of Michigan, Ann Arbor, Michigan, USA; 2University of Iowa, Iowa City, Iowa, USA; 3Section of Gastroenterology, Wake Forest University Baptist Medical Center, Winston-Salem, North Carolina, USA; 4Gastroenterology Unit, Massachusetts General Hospital, Boston, Massachusetts, USA; 5Temple University School of Medicine, Philadelphia, Pennsylvania, USA; 6University of Kansas Medical Center, Kansas City, Kansas, USA; 7 University of Buffalo, Buffalo, New York, USA; 8The SmartPill Corporation, Buffalo, New York, USA. Correspondence: Richard J. Saad, MD, Department of Internal Medicine, University of Michigan, 3912 Taubman Center, Ann Arbor, Michigan 48109, USA. E-mail: [email protected] Received 3 February 2009; accepted 16 September 2009 © 2009 by the American College of Gastroenterology

The American Journal of GASTROENTEROLOGY

1

FUNCTIONAL GI DISORDERS

ORIGINAL CONTRIBUTIONS

nature publishing group

FUNCTIONAL GI DISORDERS

2

Saad et al.

subgrouping of constipated patients on the basis of gastrointestinal and colonic transit (radio-opaque marker (ROM) studies, scintigraphy, and wireless motility capsule (WMC)) and/or anorectal and pelvic floor function (anorectal manometry, balloon expulsion testing, and defecography). Unfortunately, such specialized testing is encumbered by inconvenience, substantial cost, inconsistent availability, and lack of standardization (6). Given these limitations, specialized testing is generally reserved for patients with refractory constipation (7). Clinicians often use symptoms as surrogate markers of intestinal and colonic transit (8,9). Initial studies on clinical surrogate markers of transit were performed in individuals without bowel symptoms. Subsequent studies have assessed the accuracy of various bowel descriptors including stool form and stool frequency in predicting whole-gut transit (10,11). These later studies focused on patients with irritable bowel syndrome including those with diarrhea or constipation. In these studies, stool form was recorded using the Bristol Stool Form Scale (BSFS). The BSFS allows patients to identify their stool form using seven different images with accompanying written descriptors. Stool form was found to correlate better than stool frequency with whole-gut transit as measured by a ROM study (10,11). A follow-up study in healthy volunteers demonstrated that stool form (by BSFS) reliably predicted whole-gut transit even when altered with the use of laxatives or constipating agents (12). An additional study also suggested that stool form was a more accurate marker of intestinal transit than stool frequency in a cohort of healthy adults (13). Recent studies evaluating stool frequency as a surrogate marker have been variable with one study suggesting a good correlation between stool frequency and whole-gut transit (14) and another suggesting a poor correlation between stool frequency and colonic transit (15). Most of the published studies have been small and suffered from significant methodological limitations. Only three studies focused on patients with constipation (11,14,15) with just one of these studies assessing stool form (11). Despite the availability of only limited validation data, both stool form and stool frequency continue to be used as surrogate markers of bowel transit in research and clinical practice (16). However, the validity of stool form and stool frequency in the determination of whole-gut and colonic transit in constipation remains poorly defined. The aim of this post hoc analysis was to determine the correlation between common stool characteristics (stool form and stool frequency) and gut transit (whole-gut and colonic transit).

METHODS Data used for this post hoc analysis was derived from a large study which assessed whole-gut and colonic transit measured by two objective methods (WMC and the ROM test) in constipated and healthy adults. Details regarding the methodology and results of this study can be found elsewhere (17). We have included a summary of the methods from this study to assist in the interpretation of our results. The American Journal of GASTROENTEROLOGY

Patients

Chronically constipated and healthy subjects participating in this multicenter study were more than 18 years of age. All constipated adults fulfilled at least two of the symptoms included in the ROME II criteria for chronic functional constipation (at least 12 weeks of two or more of the following symptoms over the preceding 12 months: straining at defecation on at least a quarter of occasions, stools are lumpy/hard on at least a quarter of occasions, sensation of incomplete evacuation on at least a quarter of occasions, and/or three or fewer bowel movements a week) (18). Healthy adults were screened using the Mayo Gastrointestinal Disease screening questionnaire (19). Those with previous gastrointestinal surgery other than uncomplicated appendectomy, cholecystectomy, or cesarean section were excluded from participation in the study. Participants were required to discontinue all laxatives and other selected drugs with effects on gut motility (metoclopramide, tegaserod, erythromycin, proton pump inhibitors) for a minimum of 48 h before the first study visit. Those on stable doses of antidepressants, oral contraceptives, and antilipid agents for a minimum of 6 months were permitted to participate in this study. The study protocol was approved by the Institutional Review Boards at each of the nine participating US centers. Study protocol

All participants were required to maintain a written stool diary for 1 week before the transit testing. In this diary, participants were asked to record the time and date of every bowel movement and the stool form using the BSFS (12) (Figure 1). Following this 1-week period of self-monitoring and an overnight fast, participants presented for transit testing. All were Bristol stool chart Type 1

Separate hard lumps, like nuts (hard to pass)

Type 2

Sausage-shaped but lumpy

Type 3

Like a sausage but with cracks on its surface

Type 4

Like a sausage or snake, smooth and soft

Type 5

Soft blobs with clear-cut edges (passed easily)

Type 6

Fluffy pieces with ragged edges, a mushy stool

Type 7

Watery, no solid pieces, entirely liquid

Figure 1. Bristol Stool Form Scale (developed by Lewis and Heaton (12)) .

VOLUME 104 | XXX 2009 www.amjgastro.com

Do Stool Form and Frequency Correlate With Gut Transit

Data analysis

After collection of completed diaries from the study sites, data were compiled and analyzed including bowel movement frequency, stool form on the basis of BSFS, colonic transit time as well as whole-gut transit time measured by the WMC, and the number of retained markers 2 and 5 days after ROM ingestion for each study participant. The above-mentioned data were initially collected by the Contract Research Organization responsible for the execution of the study (Cognigen, Buffalo, NY). This data set was rechecked for accuracy through a manual review of the diaries by the study investigators blinded to the results of the transit studies. © 2009 by the American College of Gastroenterology

Calculation of stool form

Stool form was calculated for each study participant from all bowel movements recorded up to 5 days after ingestion of the WMC and the ROMs. This was reported as an average Bristol Stool Scale score represented by a numeric value ranging from 1 to 7. With the Bristol Stool Scale, lower numbers define harder stool form and vice versa (Figure 1). Calculation of stool frequency

Stool frequency was calculated for each study participant and reported as an average daily bowel movement frequency. As in the case of the stool form calculation, the stool frequency was determined by dividing the number of all reported bowel movements by the number of days the diary was maintained after ingestion of the WMC and ROMs. Assessment of colonic and whole-gut transit time by the WMC

Whole-gut transit time was defined as the time interval between the ingestion of the capsule and its exit from the body (Figure 2). Passage of the WMC outside the body was confirmed by a loss of signal and/or an abrupt drop in the temperature on the recording profile (Figure 2). Colonic transit time was defined as the time interval between the entry of the WMC into the cecum on the basis of a characteristic pH change (Figure 2) and the passage of the capsule from the body. Assessment of whole-gut transit with ROM test

The days 2 and 5 X-rays obtained on study participants from each center were collated and read by two independent investigators who were blinded to the site and to the date when the X-ray was taken. The findings of these two investigations were used for the final data analysis. A transit time was not calculated with this testing modality. Instead, delayed wholegut transit was defined as retention of >20% (five or more markers) of the ROMs 5 days after ingestion (22). There is no definition for delayed transit on the basis of a 2-day marker count, although the 2-day marker count was included in the correlation analysis (Tables 1 and 4). Statistics

The relationship of stool form and stool frequency with colonic and whole-gut transit time as measured by the WMC was assessed using Spearman’s correlation along with 95% confidence intervals based on bootstrap re-sampling. The relationship of stool form and stool frequency with whole-gut transit as defined by the ROM counts 2 and 5 days after marker ingestion was assessed in the same manner. All analyses were carried out using SAS (version 9.1, Cary, NC). The sensitivity and specificity of stool form in predicting delayed transit as measured by the WMC and ROM tests were assessed through receiver operating characteristics (ROCs) analysis. An ROC curve was generated for stool form by assessing the range of average Bristol Stool Scale values from 1 to 7 in relation The American Journal of GASTROENTEROLOGY

FUNCTIONAL GI DISORDERS

required to ingest a standardized 260 kcal nutrient bar (consisting of 17% protein, 66% carbohydrates, 2% fats, and 3% fiber) along with 50 ml of water. This was immediately followed by the ingestion of a single capsule containing 24 ROMs (Sitzmarks, Konsyl Pharmaceuticals, Fort Worth, TX) and the ingestion of the SmartPill WMC (SmartPill, Buffalo, NY). SmartPill is an indigestable capsule capable of measuring luminal pH, pressure, and temperature allowing the determination of gastric emptying (20,21) as well as intestinal, colonic, and whole-gut transit times (17). Before ingestion, the WMC was activated and calibrated using a protocol provided by the manufacturer. After ingestion of the standardized meal, WMC and ROMs, each participant was observed for a period of 6 h while fasting and then given a liquid nutrient drink (Ensure, Abbott, Abbott Park, IL). Following 2 additional hours of observation, participants were permitted to leave the test center. The data on pH, motility, and temperature were transmitted as radio frequency signals by the WMC to a receiver that was kept within a specified distance of the body for the next 5 days or until the capsule had been passed. During this period of transit testing, study participants were asked to record the time and date of every bowel movement as well as the stool form using the BSFS in a written diary. They also recorded the time and date of any additional gastrointestinal symptoms such as nausea, abdominal pain, or cramping and timing of meals, physical activity, and sleep. Restrictions during the transit testing included the avoidance of strenuous activity, abdominal crunches, prolonged aerobic activity, or alcohol consumption. Participants were instructed to consume a regular diet during the 5-day period of transit testing. There was no standardization of diet for participants before and during study participation beyond that described above on the day of capsule ingestion. Forty-eight hours after capsule ingestion, an abdominal X-ray was obtained to assess the number and distribution of ROMs. If the X-ray revealed passage of the WMC, the data recorder was collected and the capsule data were downloaded. Participants were asked to continue recording in the diaries until returning 5 days after capsule ingestion for a second abdominal X-ray (to assess the number and distribution of the ROMs).

3

4

Saad et al.

10

101

225

9

100.4

205

96.8

Body Exit

8 185

Gastric Emptying (GET)

Cecal Entry (ICJ)

93.2 7

165

86 125

5

82.4

105 4

Temperature (°F)

Pressure (mmHg)

89.6 6

145

pH 78.8

85 3 65

75.2 2

45

25

1

71.6

68

5 –5 0:07

4:48

9:36

14:24

19:12

24:00

28:48

33:36

38:24

43:12

48:00

0 52:48 55:38

64.4

Elapsed time (hr:min)

Figure 2. Data recording on SmartPill study. 130

120 Constipated Normal

110

Constipated Normal

120 110

100

100

90

90 80

WGTT

80

70

CT T

FUNCTIONAL GI DISORDERS

Patient graph 250 245

60

70 60

50

50

40

40 30

30

20

20

10

10

0

0 1

2

3

4

5

6

Average Bristol stool scale Figure 3. Scatter plot of stool form correlation with colonic transit time as measured by the wireless motility capsule. CTT, colonic transit time.

to whole-gut and colonic transit as measured by the WMC and whole-gut transit as measured by the 5-day ROM study.

1

2

3

4

5

6

Average Bristol stool scale Figure 4. Scatter plot of stool form correlation with whole-gut transit time as measured by the wireless motility capsule. WGTT, whole-gut transit time.

healthy adults. The mean 5-day ROM count was 6.8 for those with constipation compared with 0.8 for the healthy adults. Stool form and stool frequency results

RESULTS Wireless motility capsule and radio-opaque marker results

Data from 110 subjects were used in the post hoc analysis, including 46 adults with chronic constipation (91% female, mean age of 44 years, age range, 21–70 years) and 64 healthy adults (53% female, mean age of 38 years, age range, 19–72 years). The mean whole-gut transit time as measured by the WMC was 68.3 h for constipated adults compared with 33.6 h for the healthy adults. The mean colonic transit time as measured by the WMC was 59.1 h for constipated adults compared with 26.4 h for the healthy adults. The mean 2-day ROM count was 15.8 for constipated adults compared with 7.7 for the The American Journal of GASTROENTEROLOGY

Adults with constipation demonstrated an average BSFS score of 2.6 compared with a score of 3.6 for healthy adults. The mean daily bowel movement frequency was 1.0 for constipated adults compared with 1.1 for healthy adults. Correlation of stool form with transit (see Figures 3 and 4 for scatter plots)

The complete results of the correlation analysis for stool form and gut transit are detailed in Table 1. In constipated adults, stool form demonstrated a moderate correlation with whole-gut (r = − 0.61, P < 0.0001) and colonic transit ( − 0.62, P < 0.0001) times as measured by the WMC. In the constipated VOLUME 104 | XXX 2009 www.amjgastro.com

Do Stool Form and Frequency Correlate With Gut Transit

5

Overall

Constipated

Healthy

Whole-gut transit time by wireless motility capsule

− 0.47; 95% CI: − 0.63 to − 0.30; P < 0.0001

− 0.61; 95% CI: − 0.77 to − 0.37; P < 0.0001

− 0.13; 95% CI: − 0.36 to 0.14; P = 0.31

Colonic transit time by wireless motility capsule

− 0.48; 95% CI: − 0.63 to − 0.30; P < 0.0001

− 0.62; 95% CI: − 0.78 to − 0.40; P < 0.0001

− 0.10; 95% CI: − 0.33 to 0.16; P = 0.44

Day 2 radio-opaque marker count

− 0.50; 95% CI: − 0.65 to − 0.33; P < 0.0001

− 0.61; 95% CI: − 0.78 to − 0.35; P < 0.0001

− 0.10; 95% CI: − 0.39 to 0.09; P = 0.44

Day 5 radio-opaque marker count

− 0.36; 95% CI: − 0.53 to − 0.16; P < 0.0001

− 0.45; 95% CI: − 0.68 to − 0.19; P = 0.0016

0.02; 95% CI: − 0.25 to 0.27; P = 0.84

CI, confidence interval.

Table 2. Accuracy of stool form (BSFS value from single bowel movement) in predicting delayed transit CTT by wireless motility capsule Stool form (value on BSFS)

WGTT by wireless motility capsule

WGT by 5-day ROM study

Sensitivity (%)

Specificity (%)

Sensitivity (%)

Specificity (%)

Sensitivity (%)

Specificity (%)

1

50

92

55

92

40

89

2

86

74

85

72

70

69

3

95

47

95

46

85

43

4

100

11

100

11

95

10

BSFS, Bristol Stool Form Scale; CTT, colonic transit time; ROM, radio-opaque marker; WGT, whole-gut transit; WGTT, whole-gut transit time.

Table 3. Accuracy of stool form (average BSFS value over 2–5 days) in predicting delayed transit CTT by wireless motility capsule Stool form (value on BSFS)

WGTT by wireless motility capsule

WGT by 5-day ROM study

Sensitivity (%)

Specificity (%)

Sensitivity (%)

Specificity (%)

Sensitivity (%)

Specificity (%)

1

14

100

15

100

10

99

2

64

90

65

89

55

87

2.5

82

83

85

82

80

81

3

86

72

90

71

80

69

4

100

22

100

21

95

20

BSFS, Bristol Stool Form Scale; CTT, colonic transit time; ROM, radio-opaque marker; WGT, whole-gut transit; WGTT, whole-gut transit time.

women, the correlation coefficient for stool form was − 0.64 (P < 0.0001) with whole-gut transit and − 0.66 (P < 0.0001) for colonic transit. For the four constipated men in the study, the correlation coefficients were − 0.4 and − 0.4, respectively. In contrast, for healthy adults the correlation between stool form and gut transit was weak yielding a correlation coefficient of − 0.13 for whole-gut transit and − 0.10 for colonic transit. In healthy women, the correlation coefficient was − 0.14 for whole-gut transit and − 0.08 for colonic transit. For healthy men, the correlation coefficients were − 0.19 and − 0.19, respectively. In constipated adults, a moderate correlation was also found between stool form and transit as assessed by the ROM test, yielding a correlation coefficient of − 0.45 (P = 0.0016) with © 2009 by the American College of Gastroenterology

the 5-day ROM count. The correlation coefficient was − 0.45 (P < 0.003) in constipated women and − 0.4 in the constipated men. No correlation between stool form and the 5-day ROM count was observed in healthy volunteers. Sensitivity and specificity of stool form for delayed transit

The results of the ROC analysis for stool form in relation to bowel transit as measured by the WMC and 5-day ROM study are detailed in Tables 2 and 3. This analysis was performed on two measures of stool form, one representing the average Bristol stool form score for all bowel movements recorded by subjects during transit testing (Table 3), the other representing the first recorded bowel movement during transit The American Journal of GASTROENTEROLOGY

FUNCTIONAL GI DISORDERS

Table 1. Spearman’s correlation of stool form with measured transit

6

Saad et al.

FUNCTIONAL GI DISORDERS

Table 4. Spearman’s correlation of stool frequency with measured transit Overall

Constipated

Healthy

Whole-gut transit time by wireless motility capsule

− 0.16; 95% CI: − 0.36 to 0.05; P = 0.1

− 0.02; 95% CI: − 0.34 to 0.27; P = 0.89

− 0.25; 95% CI: − 0.48 to 0.01; P = 0.04

Colonic transit time by wireless motility capsule

− 0.17; 95% CI: − 0.37 to 0.02; P = 0.08

− 0.03; 95% CI: − 0.33 to 0.26; P = 0.85

− 0.26; 95% CI: − 0.49 to 0.00; P = 0.04

Day 2 radio-opaque marker count

− 0.08; 95% CI: − 0.28 to 0.11; P = 0.40

0.04; 95% CI: − 0.28 to 0.35; P = 0.79

− 0.13; 95% CI: − 0.36 to 0.12; P = 0.30

Day 5 radio-opaque marker count

− 0.09; 95% CI: − 0.29 to 0.11; P = 0.34

0.04; 95% CI: − 0.27 to 0.35; P = 0.79

− 0.24; 95% CI: − 0.41 to − 0.06; P = 0.05

CI, confidence interval.

Table 5. Spearman’s correlation of stool frequency with measured transit on the basis of weekly bowel movement count

Less than three bowel movements a week

Less than four bowel movements a week

Less than five bowel movements a week

Less than six bowel movements a week

Whole-gut transit by wireless motility capsule

Colonic transit by wireless motility capsule

Day 5 radio-opaque marker count

0.79

0.79

0.64

n =11,

n =11,

n =11,

− 0.29

− 0.21

− 0.008

n =18,

n =18,

n =18,

− 0.08,

− 0.11

0.05

n =31,

n =31,

n =31,

− 0.05

− 0.07

− 0.03

n = 33

n = 33

n = 33

testing (Table 2). Data were combined from the entire study population of 110 subjects for these ROC analyses. An average stool form cutoff value of 2.5 most accurately predicted delayed transit as measured by the WMC and ROM studies. A stool form cutoff value of 2.5 yielded a sensitivity of 82% and specificity of 83% in relation to measured colonic transit by the WMC, sensitivity of 85% and specificity of 82% in relation to measured whole-gut transit by the WMC, and sensitivity of 80% as well as specificity of 81% in relation to the 5-day ROM study. For the first bowel movement reported during these motility studies, a Bristol stool form cutoff value of 2 was most accurate. This value yielded a sensitivity of 86% and specificity of 74% in relation to measured colonic transit by the WMC, sensitivity of 85% and specificity of 72% in relation to measured whole-gut transit by the WMC, and sensitivity of 70% as well as specificity of 69% in relation to the 5-day ROM study. The American Journal of GASTROENTEROLOGY

Correlation of stool frequency with transit

The complete results of the correlation analysis for stool frequency and gut transit are detailed in Table 4. In constipated adults, there was no correlation between stool frequency and either whole-gut (r = − 0.02) or colonic transit times (r = − 0.03) as measured by the WMC. The correlation between stool frequency and whole-gut (r = − 0.25) or colonic (r = − 0.26) transit as measured by the WMC was also poor in healthy adults. The correlation between stool frequency and whole-gut transit as assessed by the ROM test was weak in both constipated and healthy adults with correlation coefficients of 0.04 and − 0.24, respectively. Correlation between reduced stool frequency and transit

We were interested in determining whether abnormally reduced stool frequency (those reporting fewer than three bowel movements a week (23,24)) provided a better correlation with transit than that observed for the overall constipation cohort. For this analysis, we stratified the constipation cohort on the basis of weekly stool frequency. Correlation coefficients for stool frequency with whole-gut and colonic transit were calculated for four subgroups of constipated patients: those reporting fewer than three bowel movements per week, those reporting fewer than four bowel movements per week, those reporting fewer than five bowel movements per week, and those reporting fewer than six bowel movements per week. The results of this analysis are detailed in Table 5. The correlation coefficients with whole-gut and colonic transit as measured by the WMC and the 5-day ROM test were poor in all constipation subgroups regardless of weekly stool frequency.

DISCUSSION Because of the lack of availability, inconvenience, and cost of tests which allow the formal measurement of whole-gut and colonic transit, physicians typically use information gathered during the patient interview to assess constipation severity and to determine a treatment strategy. The extent to which constipation-associated symptoms, such as stool form and stool frequency, reflect quantifiable abnormalities in gastrointestinal VOLUME 104 | XXX 2009 www.amjgastro.com

or colonic functional measures such as transit remains poorly validated. This study provides the largest available data set addressing the correlation between the clinical parameters of stool form and stool frequency with whole-gut and colonic transit in both constipated and healthy adults. It is also the first study in which these clinical parameters were compared with the results of two simultaneously administered objective studies of whole-gut and colonic transit. Most of the available data addressing the correlation between symptoms and transit have been derived from studies in healthy adults. Our study is the first to have included a substantial number of adults with chronic constipation. Only one other small study has included patients with constipationrelated complaints (11). In this study, Heaton and O’Donnell (11) assessed symptoms and used a modified ROM protocol to assess whole-gut transit in 47 patients with the irritable bowel syndrome, only a subgroup of which had constipation-related complaints. Similar to this study, they found a modest correlation between stool form and whole-gut transit (r = − 0.57) but no correlation between stool frequency and transit (r = 0.31). Our ROC analysis provides clinically relevant information regarding the use of stool form in predicting delayed transit. The sensitivity and specificity for various Bristol stool form values suggest that stool form either reported on a single bowel movement or averaged from bowel movements reported in a diary over several days is reasonably accurate in predicting whether an individual has delayed vs. normal whole-gut or colonic transit. It is worth pointing out that the sensitivity and specificity calculations pertain to a dichotomous outcome—delayed or normal transit. In other words, determination of stool form provides a qualitative rather than quantitative assessment of transit. From a clinical perspective, this means that stool form can be used to predict delayed vs. normal transit, but cannot reliably gauge the severity of delayed transit. In contrast to earlier studies by Lewis and Heaton (12) and Degan and Phillips (13), our analysis demonstrated a poor correlation (r = − 0.13) between stool form and gut transit in healthy adults. In 66 healthy women, Lewis and Heaton (12) reported a modest correlation between stool form and wholegut transit as measured by a modified ROM study with an r = − 0.54. In another study by Degen and Phillips (13), 32 healthy adults (63% women) were evaluated with scintigraphy and a ROM study. A ROM test using the Metcalf protocol (25) allowed determination of segmental and total colonic transit time. Using regression analysis, the linear relationship between stool form and colonic transit in this study was y = 4.7 − 0.04x, P < 0.001 (13). There are a number of potential explanations for the differences in our results and those from these previous studies. Inherent differences in study populations and testing modalities likely provide explanations for some of the observed discrepancies. For example, the average whole-gut transit time for healthy adults in our study as measured by WMC was 33.6 h compared with 62.7 h by ROMs in the Lewis and Heaton study. This striking difference in whole-gut transit time results may © 2009 by the American College of Gastroenterology

also in part be related to the differences in the techniques used to measure gut transit. Consistent with the published literature (11,15,26), our analysis found a poor correlation between stool frequency and whole-gut or colonic transit time as measured by the WMC and ROM testing in constipated and healthy adults. The study by Heaton and O’Donnell (11) reported a poor correlation (r = 0.31) between stool frequency and whole-gut transit in 47 irritable bowel syndrome patients. Chaussade et al. (15) reported no correlation between stool frequency and colonic transit as measured by a ROM study in 91 (71% female) adults with constipation, although no specific correlation coefficients were reported. Wyman et al. (26) also reported a poor correlation between stool frequency and transit as measured by a ROM study in a cohort of 10 healthy women. The poor correlation values we observed in both constipated and healthy adults for all objective measures of colonic and whole-gut transit bolster these earlier reported observations. Given the poor correlation between stool frequency and gut transit in the constipated and healthy populations, we performed an analysis to determine whether abnormally reduced stool frequency (clinically defined as fewer than three bowel movements a week (23,24)) might provide a better clinical surrogate marker of slow transit. This analysis confirmed a poor correlation between stool frequency and transit even among those patients who reported fewer than three bowel movements per week. Correlations were poor for all objective measures of whole-gut and colonic transit by both the WMC and ROM test. The only published comparable data come from a report by Glia et al. (14) who studied 134 constipated patients with a 5-day ROM study. Those with slow transit on the marker study were more likely to report two or fewer bowel movements per week than those with normal transit (84 vs. 46%, P < 0.05) (14). It is difficult to directly compare the findings of this study with those of the Glia study given the differences in study design and study objectives. Collectively, these studies suggest that although adults with slow-transit constipation are more likely to report an abnormally low stool count, stool frequency remains a poor clinical surrogate for objectively measured whole-gut and colonic transit time. This study had several limitations. We enrolled only constipated or healthy adults. Those with diarrheal symptoms were excluded. As such, we are unable to make any statements regarding the correlation between stool form or frequency and transit in patients with diarrhea. Furthermore, there were only four constipated men included in the analysis limiting our ability to comment with any certainty on the correlation between stool form and measured transit in constipated men. Although the correlation between stool form and measured transit appeared moderate with a correlation coefficient of − 0.4 for both wholegut and colonic transit time in the constipated men, no definitive conclusions can be drawn from such a small study set. In addition, our analysis evaluated chronic constipation in general terms. Chronic constipation represents a heterogeneous disorder with a multifaceted pathogenesis (27). As such, constipation The American Journal of GASTROENTEROLOGY

7

FUNCTIONAL GI DISORDERS

Do Stool Form and Frequency Correlate With Gut Transit

FUNCTIONAL GI DISORDERS

8

Saad et al.

can be subtyped on the basis of the prevailing underlying pathophysiological mechanisms responsible for symptoms. Such subgroups include slow-transit constipation, dyssynergic defecation, and normal transit constipation (5). Our post hoc analysis was not designed or powered to address the correlation of stool form or stool frequency with bowel transit in these subtypes of constipation. The correlation of clinical parameters including stool form and stool frequency with that of gut transit is likely to vary depending on the underlying pathophysiology of the constipation. There was also no standardization, restriction, or supplementation of any subjects’ diets before and during the transit testing. The investigators’ primary objectives in the original study were to assess transit times under real physiological conditions and not to artificially change subjects’ diets. As this is a post hoc analysis, we had no control over this issue of diet standardization. It is therefore plausible that differences in diet among the subjects could have affected the relationship between bowel symptoms and measured transit times. An argument can also be made regarding the clinical and scientific validity of using averaged stool form data obtained over several days’ time. Because stool characteristics vary between and even within individuals, particularly in constipated persons, we felt that assessing these parameters over a period of days would provide a better estimate of the subject’s actual stool characteristics than would collecting data at a single point in time. Regarding stool form, we performed an ROC analysis using the average stool form value obtained over several days and also for the form of the first reported bowel movement after ingestion of the WMC. Although averaged stool form scores and a single stool form score yielded a slightly different stool form value which predicted the presence of delayed gut transit, both provided a reasonable qualitative assessment of gut transit. In summary, stool form as measured by the BSFS provides a reasonably accurate qualitative assessment of whole-gut and colonic transit. Stool form and measured whole-gut as well as colonic transit time were better correlated in constipated than in healthy adults. Stool frequency correlated poorly with all transit measurements in constipated and healthy adults. On the basis of our results, we suggest that stool form but not stool frequency can be used as a simple office-based surrogate for whole-gut and colonic transit. Unfortunately, stool form provides only a qualitative assessment of transit. As such, when a precise estimate of whole-gut or colonic transit is necessary, formal evaluation with an objective testing modality such as the WMC, scintigraphy, or a ROM test should be pursued. CONFLICT OF INTEREST

Guarantor of the article: Richard J. Saad, MD. Specific author contributions: All authors participated in the planning and conduct of this clinical trial. Saad, Semler, Wilding, and Chey analyzed the data. Saad and Chey prepared and are responsible for the full contents of the manuscript. Rao and Parkman assisted with editing and revision of the manuscript. All authors reviewed and approved the final manuscript. The American Journal of GASTROENTEROLOGY

Financial support: The clinical trial of the WMC from which this post hoc analysis was performed was supported by the SmartPill Corporation, Buffalo, NY. Potential competing interests: Rao, Koch, Kuo, Parkman, McCullum, Sitrin, and Chey serve as speakers, consultants, or advisory board members for the SmartPill. Saad has received research funding from the SmartPill. Wilding serves as a consultant to the SmartPill. Semler is an employee of the SmartPill and owns stock in the SmartPill.

Study Highlights WHAT IS CURRENT KNOWLEDGE Constipation is a symptom-based disorder. Symptoms, including stool form and stool frequency, are used as surrogate markers of bowel transit. Stool form is believed to be more accurate than stool frequency in predicting whole-gut transit. WHAT IS NEW HERE Stool form, expressed as a value on the Bristol Stool Scale, provides an accurate qualitative assessment of whole-gut and colonic transit in adults. A modest correlation exists between stool form and measured whole-gut as well as colonic transit time in chronic constipation. Stool frequency correlates poorly with both measured whole-gut and colonic transit time. Neither stool form nor stool frequency provides a reliable measure of whole-gut or colonic transit time. When such information is needed, it should be obtained using an objective testing modality.

3 3 3 3 3 3 3

REFERENCES 1. Longstreth GF, Thompson WG, Chey WD et al. Functional bowel disorders. Gastroenterology 2006;130:1480–91. 2. American College of Gastroenterology Chronic Constipation Task Force. An evidence-based approach to the management of chronic constipation in North America. Am J Gastroenterol 2005;100 (Suppl 1): S1–4. 3. Andromanakos N, Skandalakis P, Troupis T et al. Constipation of anorectal outlet obstruction: pathophysiology, evaluation and management. J Gastroenterol Hepatol 2006;21:638–46. 4. Bharucha AE. Update of tests of colon and rectal structure and function. J Clin Gastroenterol 2006;40:96–103. 5. Rao SS. Constipation: evaluation and treatment of colonic and anorectal motility disorders. Gastroenterol Clin North Am 2007;36:687–711, x. 6. Rao SS, Ozturk R, Laine L. Clinical utility of diagnostic tests for constipation in adults: a systematic review. Am J Gastroenterol 2005;100:1605–15. 7. Locke GR III, Pemberton JH, Phillips SF. American Gastroenterological Association Medical Position Statement: guidelines on constipation. Gastroenterology 2000;119:1761–6. 8. Davies GJ, Crowder M, Reid B et al. Bowel function measurements of individuals with different eating patterns. Gut 1986;27:164–9. 9. Probert CJ, Emmett PM, Heaton KW. Intestinal transit time in the population calculated from self made observations of defecation. J Epidemiol Community Health 1993;47:331–3. 10. O’Donnell LJ, Virjee J, Heaton KW. Detection of pseudodiarrhoea by simple clinical assessment of intestinal transit rate. Br Med J (Clinical research ed) 1990;300:439–40. 11. Heaton KW, O’Donnell LJ. An office guide to whole-gut transit time. Patients’ recollection of their stool form. J Clin Gastroenterol 1994;19:28–30. 12. Lewis SJ, Heaton KW. Stool form scale as a useful guide to intestinal transit time. Scand J Gastroenterol 1997;32:920–4.

VOLUME 104 | XXX 2009 www.amjgastro.com

13. Degen LP, Phillips SF. How well does stool form reflect colonic transit? Gut 1996;39:109–13. 14. Glia A, Lindberg G, Nilsson LH et al. Clinical value of symptom assessment in patients with constipation. Dis Colon Rectum 1999;42:1401–8; discussion 1408–10. 15. Chaussade S, Khyari A, Roche H et al. Determination of total and segmental colonic transit time in constipated patients. Results in 91 patients with a new simplified method. Dig Dis Sci 1989;34:1168–72. 16. Choung RS, Locke GR III, Zinsmeister AR et al. Epidemiology of slow and fast colonic transit using a scale of stool form in a community. Aliment Pharmacol Ther 2007;26:1043–50. 17. Rao SS, Kuo B, McCallum RW et al. Investigation of colonic and whole-gut transit with wireless motility capsule and radiopaque markers in constipation. Clin Gastroenterol Hepatol 2009;7:537–44. 18. Thompson WG, Longstreth GF, Drossman DA et al. Functional bowel disorders and functional abdominal pain. Gut 1999;45 (Suppl 2): II43–7. 19. Locke GR, Talley NJ, Weaver AL et al. A new questionnaire for gastroesophageal reflux disease. Mayo Clin Proc 1994;69:539–47.

© 2009 by the American College of Gastroenterology

20. Cassilly D, Kantor S, Knight LC et al. Gastric emptying of a non-digestible solid: assessment with simultaneous SmartPill pH and pressure capsule, antroduodenal manometry, gastric emptying scintigraphy. Neurogastroenterol Motil 2008;20:311–9. 21. Kuo B, McCallum RW, Koch KL et al. Comparison of gastric emptying of a nondigestible capsule to a radio-labelled meal in healthy and gastroparetic subjects. Aliment Pharmacol Ther 2008;27:186–96. 22. Evans RC, Kamm MA, Hinton JM et al. The normal range and a simple diagram for recording whole gut transit time. Int J Colorectal Dis 1992;7:15–7. 23. Connell AM, Hilton C, Irvine G et al. Variation of bowel habit in two population samples. Br Medi J 1965;2:1095–9. 24. Drossman DA, Sandler RS, McKee DC et al. Bowel patterns among subjects not seeking health care. Use of a questionnaire to identify a population with bowel dysfunction. Gastroenterology 1982;83:529–34. 25. Metcalf AM, Phillips SF, Zinsmeister AR et al. Simplified assessment of segmental colonic transit. Gastroenterology 1987;92:40–7. 26. Wyman JB, Heaton KW, Manning AP et al. Variability of colonic function in healthy subjects. Gut 1978;19:146–50. 27. Lembo A, Camilleri M. Chronic constipation. N Engl J Med 2003;349:1360–8.

The American Journal of GASTROENTEROLOGY

9

FUNCTIONAL GI DISORDERS

Do Stool Form and Frequency Correlate With Gut Transit