ORIGINAL CONTRIBUTIONS
nature publishing group
1001
The Relationship Among Perceived Stress, Symptoms, and Inflammation in Persons With Inflammatory Bowel Disease Laura E. Targownik, MD, MSHS1, Kathryn A. Sexton, PhD2, Matthew T. Bernstein, BA (Hons)2, Brooke Beatie, BA (Hons)2, Michael Sargent, BSc1, John R. Walker, PhD2 and Lesley A. Graff, PhD2 OBJECTIVES:
Previous studies have demonstrated that stress is associated with increased disease activity in individuals with inflammatory bowel disease (IBD). The association between perceived stress and gastrointestinal inflammation is not well described.
METHODS:
Participants were recruited from a population-based registry of individuals with known IBD. Symptomatic disease activity was assessed using validated clinical indices: the Manitoba IBD Index (MIBDI) and Harvey Bradshaw Index (HBI) for Crohn’s disease (CD), and Powell Tuck Index (PTI) for ulcerative colitis (UC). Perceived stress was measured using Cohen’s Perceived Stress Scale (CPSS). Intestinal inflammation was determined through measurement of fecal calprotectin (FCAL), with a level exceeding 250 μ g/g indicating significant inflammation. Logistic regressions were used to evaluate the association between intestinal inflammation, perceived stress, and disease activity.
RESULTS:
Of the 478 participants with completed surveys and stool samples, perceived stress was associated with symptomatic activity (MIBDI) for both CD and UC (1.07 per 1-point increase on the CPSS, 95% confidence interval (CI) 1.03–1.10 and 1.03–1.11, respectively). There was no significant association between perceived stress and intestinal inflammation for either CD or UC. Active symptoms (MIBDI ≤3) were associated with intestinal inflammation in UC (odds ratio (OR) 3.94, 95% CI 1.65–9.43), but not in CD (OR 0.98, 95% CI 0.51–1.88).
CONCLUSIONS: Symptomatic disease activity was unrelated to intestinal inflammation in CD and only weakly
associated in UC. Although there was a strong relationship between perceived stress and gastrointestinal symptoms, perceived stress was unrelated to concurrent intestinal inflammation. Longitudinal investigation is required to determine the directionality of the relationship between perceived stress, inflammation, and symptoms in IBD. AM J Gastroenterol 2015; 110:1001–1012; doi:10.1038/ajg.2015.147; published online 16 June 2015
INTRODUCTION Inflammatory bowel disease (IBD), including Crohn’s disease (CD) and ulcerative colitis (UC), is a chronic condition characterized by episodes of gastrointestinal (GI) symptoms. Inflammation of the GI tract can lead to patient symptoms in IBD, and the development of complications including intestinal perforations, strictures, and fistulas. The goal of medical management of IBD is to decrease symptom burden and prevent complications
by decreasing the intensity of intestinal inflammation. Therapies that result in complete control of intestinal inflammation are associated with favorable long-term outcomes for individuals with IBD, both in terms of reducing symptoms and of decreasing the incidence of complications (1–4). However, individuals with IBD may also experience GI symptoms independent of IBD-related inflammation (5–7). Notably, IBD and functional bowel disorders may coexist and
1
Department of Internal Medicine, University of Manitoba, Winnipeg, Manitoba, Canada; 2Department of Clinical Health Psychology, University of Manitoba, Winnipeg, Manitoba, Canada. Correspondence: Laura E. Targownik, MD, MSHS, Department of Internal Medicine, University of Manitoba, 805G-715 McDermot Avenue, Winnipeg, Manitoba R3E 3P4, Canada. E-mail: [email protected] Received 25 September 2014; accepted 24 March 2015
© 2015 by the American College of Gastroenterology
The American Journal of GASTROENTEROLOGY
INFLAMMATORY BOWEL DISEASE
CMErelated editorial on page x see
INFLAMMATORY BOWEL DISEASE
1002
Targownik et al.
they share similar symptoms (7). Previous inflammation may induce changes in structure and function that result in symptoms in the absence of ongoing inflammation. However, it is unlikely the anti-inflammatory regimens used to treat IBD would be effective in treating symptoms when inflammation is not present. It is also recognized that individuals with IBD who are experiencing symptoms are more likely to have higher levels of perceived stress (8–10). Perceived stress, which takes into account the perceived controllability, manageability, and personal impact of situations, has a strong and consistent association with concurrent and subsequent IBD symptoms (11). However, it is unclear whether stress induces increased levels of gut inflammation, or whether individuals with higher stress experience more intense symptoms independent of gut inflammation. Conversely, individuals with gut inflammation or a high symptom burden may have higher levels of perceived stress. Although the gold standard for determining the presence of active inflammation is through performance of endoscopy with histologic sampling (12), endoscopy is generally not immediately accessible for individuals who present to an outpatient clinic with troublesome GI symptoms. Standard biochemical tests such as serum C-reactive protein and erythrocyte sedimentation rate may suggest the presence of inflammation (13,14). However, these tests have limited sensitivity and specificity, and therefore many clinicians elect to treat symptoms without definitively confirming the presence of GI inflammation. A recent survey of gastroenterologists reported that the majority use symptomatic activity rather than the presence of intestinal inflammation (15) to guide therapy. Treating patient symptoms without confirming the presence and/or absence of inflammation may lead to unnecessary exposure to therapies that may have significant side effects as well as high costs. Recently, there has been increasing evidence that the level of fecal calprotectin (FCAL) is very sensitive and specific for the presence of intestinal inflammation in both CD and UC (16,17). FCAL is released from the breakdown of neutrophils in the intestines, and is not elevated in the absence of gut inflammation. It has been shown to be useful in differentiating symptomatic patients with IBD from those with functional GI disorders such as irritable bowel syndrome (18,19). FCAL levels are also well correlated with the degree of endoscopic and histological inflammation seen at endoscopy (19). However, the relationships among inflammatory activity, symptoms in IBD, and perceived stress are not well understood. Therefore, using a cross-sectional design and a population-based sample of individuals with IBD, we aimed to characterize the relationship among stress, symptom activity, and inflammatory activity, the latter determined through assessment of FCAL.
METHODS Sample
Participants were recruited from the University of Manitoba Research Registry (20). The Registry was developed using a The American Journal of GASTROENTEROLOGY
well-established administrative definition of IBD to identify all individuals in Manitoba with IBD. In 2000, and again in 2008, all individuals who met this administrative definition for IBD in the provincial single-insurer health database were contacted to invite them to be part of the Research Registry, with a response rate of ∼ 50%. Inclusion in the Registry is voluntary, and participants are not compensated for agreeing to be included. Individuals in the Registry agree to be contacted about various research initiatives, but are not obligated to participate. The demographic composition of the Research Registry is comparable to the broad population of individuals with IBD in Manitoba (21). At the time of recruitment, the Registry included 1,958 living individuals aged 250 μ g/g). Perceived stress was entered as a continuous variable, such that odds ratios (ORs) reflect the risk of increased IBD symptoms or intestinal inflammation associated with a 1-point change in level of perceived stress on the CPSS. We opted to employ the CPSS scores in this way to take advantage of the increased statistical power afforded by the use of a continuous scale. The P values of 2 units/day)
10 (3.8%)
10 (4.7%)
Prednisone
16 (6.1%)
10 (4.6%)
Biologic agent
37 (14.3%)
10 (4.7%)
3.9±1.5
4.2±1.6
93 (36.0%)
56 (26.2%)
5.6±4.1
4.3±3.6
132 (54.8%)
71 (39.0%)
20.9±8.7
19.4±9.4
Medication use in previous 3 months
Disease activity symptom scores MIBDI, mean (s.d.) MIBDI, proportion active (symptoms ≥ a few days every other week) n (%) Harvey Bradshaw/Powell Tuck Index, mean±s.d. Harvey Bradshaw/Powell Tuck, proportion active (≥5) n(%) Perceived stress Cohen Perceived Stress Score (mean±s.d.) IBD, inflammatory bowel disease; MIBDI, Manitoba IBD Index.
with lengthy disease (i.e., ≥20 years) had no current inflammation (63.7% vs. 48.0%, P=0.013). Considering the participants with UC, 34.5% overall were found to have elevated FCAL levels. Comparing those with normal and elevated FCAL, a considerably higher proportion of those with symptomatically active disease as measured by the MIBDI had an elevated FCAL level (42.7% vs. 17.3%, P0.2
N (%) ≥45
Mean±s.d.
127 (79.9%)
76 (74.5%)
>0.2
122 (85.9%)
61 (77.3%)
0.110
Female
113 (71.1%)
66 (64.7%)
>0.2
77 (54.4%)
47 (62.7%)
>0.2
7 (4.5%)
9 (9.0%)
0.153
5 (3.6%)
5 (6.8%)
>0.2
21 (13.4%)
16 (15.7%)
>0.2
5 (3.6%)
5 (6.8%)
>0.2
23.5±10.4
22.4±10.8
>0.2
22.2±11.8
18.0±9.8
0.009
Medication use in past 3 months Prednisone Biologic Duration of disease Mean years±s.d. N (%) with ≥20 years of duration
100 (63.7%)
49 (48.0%)
0.013
76 (54.7%)
27 (36.5%)
0.011
Current smoker
15 (9.5%)
15 (14.7%)
>0.2
9 (6.4%)
5 (6.7%)
>0.2
Alcohol use (≥2 units/day)
5 (3.1%)
5 (5.0%)
>0.2
5 (3.6%)
5 (6.7%)
>0.2
91 (57.6%)
55 (54.5%)
>0.2
10 (7.0%)
6 (8.1%)
>0.2
History of IBD-related surgery
FCAL, fecal calprotectin (μ g per g stool); HBI, Harvey Bradshaw Index; IBD, inflammatory bowel disease; IQR, interquartile range; MIBDI, Manitoba IBD Index; PTI, Powell Tuck Index. Proportions are reported as percentages of those with active inflammation or with low inflammation who reported a high level of symptoms. MIBDI: reports of symptoms occurring sometimes (a few days every other week), often (some days each week), or all the time (every day) were defined as actively symptomatic; higher scores on the MIBDI indicate less frequent symptoms. a N=258 in CD and n=214 in UC. Harvey Bradshaw Clinical Index for Crohn’s disease; scores of ≥5 were defined as actively symptomatic; higher scores indicated more frequent/severe symptoms. b N=241 for % comparisons and n=229 for mean comparisons. Powell Tuck Clinical Index for ulcerative colitis; scores of ≥5 were defined as actively symptomatic; higher scores indicate more frequent/severe symptoms. c N=182 for % comparisons and n=181 for mean comparisons. Perceived stress was assessed using Cohen’s Perceived Stress Scale; higher scores indicate higher perceived stress. d N=260 in CD and n=215 in UC.
Multivariate logistic regression was used to evaluate the relationship of several personal and disease characteristics with active symptomatic disease for respondents with CD and UC, and the results are shown in Tables 5 and 6 respectively. For the CD group, there was a greater likelihood of active symptomatic disease, based on the HBI, given a recent use of prednisone, a history of IBD surgery, current smoking, and higher concurrent perceived stress, controlling for the other factors. Active symptomatic disease based on the MIBDI was significantly more likely if a biologic medication had been used in recent months and with higher concurrent perceived stress. Intestinal inflammation (elevated FCAL) was not associated with active symptomatic disease, as measured by either the MIBDI or HBI. For the UC group, previous IBD surgery, © 2015 by the American College of Gastroenterology
current elevated FCAL levels, and current higher perceived stress levels were all significantly associated with active symptomatic disease across both measures (MIBDI and PTI), controlling for other factors. Finally, the relationship of these personal and disease characteristics with elevated fecal calprotectin was evaluated separately for CD and UC (Table 7) using logistic regression. Considering the results for the CD group, elevated FCAL level was not significantly associated with most of the variables in the analysis, including clinical factors such as age at diagnosis, history of GI surgery, current biologic use, current prednisone use, smoking status, or current level of perceived stress. Elevated FCAL was associated with shorter disease duration (i.e.,
nature publishing group
1001
The Relationship Among Perceived Stress, Symptoms, and Inflammation in Persons With Inflammatory Bowel Disease Laura E. Targownik, MD, MSHS1, Kathryn A. Sexton, PhD2, Matthew T. Bernstein, BA (Hons)2, Brooke Beatie, BA (Hons)2, Michael Sargent, BSc1, John R. Walker, PhD2 and Lesley A. Graff, PhD2 OBJECTIVES:
Previous studies have demonstrated that stress is associated with increased disease activity in individuals with inflammatory bowel disease (IBD). The association between perceived stress and gastrointestinal inflammation is not well described.
METHODS:
Participants were recruited from a population-based registry of individuals with known IBD. Symptomatic disease activity was assessed using validated clinical indices: the Manitoba IBD Index (MIBDI) and Harvey Bradshaw Index (HBI) for Crohn’s disease (CD), and Powell Tuck Index (PTI) for ulcerative colitis (UC). Perceived stress was measured using Cohen’s Perceived Stress Scale (CPSS). Intestinal inflammation was determined through measurement of fecal calprotectin (FCAL), with a level exceeding 250 μ g/g indicating significant inflammation. Logistic regressions were used to evaluate the association between intestinal inflammation, perceived stress, and disease activity.
RESULTS:
Of the 478 participants with completed surveys and stool samples, perceived stress was associated with symptomatic activity (MIBDI) for both CD and UC (1.07 per 1-point increase on the CPSS, 95% confidence interval (CI) 1.03–1.10 and 1.03–1.11, respectively). There was no significant association between perceived stress and intestinal inflammation for either CD or UC. Active symptoms (MIBDI ≤3) were associated with intestinal inflammation in UC (odds ratio (OR) 3.94, 95% CI 1.65–9.43), but not in CD (OR 0.98, 95% CI 0.51–1.88).
CONCLUSIONS: Symptomatic disease activity was unrelated to intestinal inflammation in CD and only weakly
associated in UC. Although there was a strong relationship between perceived stress and gastrointestinal symptoms, perceived stress was unrelated to concurrent intestinal inflammation. Longitudinal investigation is required to determine the directionality of the relationship between perceived stress, inflammation, and symptoms in IBD. AM J Gastroenterol 2015; 110:1001–1012; doi:10.1038/ajg.2015.147; published online 16 June 2015
INTRODUCTION Inflammatory bowel disease (IBD), including Crohn’s disease (CD) and ulcerative colitis (UC), is a chronic condition characterized by episodes of gastrointestinal (GI) symptoms. Inflammation of the GI tract can lead to patient symptoms in IBD, and the development of complications including intestinal perforations, strictures, and fistulas. The goal of medical management of IBD is to decrease symptom burden and prevent complications
by decreasing the intensity of intestinal inflammation. Therapies that result in complete control of intestinal inflammation are associated with favorable long-term outcomes for individuals with IBD, both in terms of reducing symptoms and of decreasing the incidence of complications (1–4). However, individuals with IBD may also experience GI symptoms independent of IBD-related inflammation (5–7). Notably, IBD and functional bowel disorders may coexist and
1
Department of Internal Medicine, University of Manitoba, Winnipeg, Manitoba, Canada; 2Department of Clinical Health Psychology, University of Manitoba, Winnipeg, Manitoba, Canada. Correspondence: Laura E. Targownik, MD, MSHS, Department of Internal Medicine, University of Manitoba, 805G-715 McDermot Avenue, Winnipeg, Manitoba R3E 3P4, Canada. E-mail: [email protected] Received 25 September 2014; accepted 24 March 2015
© 2015 by the American College of Gastroenterology
The American Journal of GASTROENTEROLOGY
INFLAMMATORY BOWEL DISEASE
CMErelated editorial on page x see
INFLAMMATORY BOWEL DISEASE
1002
Targownik et al.
they share similar symptoms (7). Previous inflammation may induce changes in structure and function that result in symptoms in the absence of ongoing inflammation. However, it is unlikely the anti-inflammatory regimens used to treat IBD would be effective in treating symptoms when inflammation is not present. It is also recognized that individuals with IBD who are experiencing symptoms are more likely to have higher levels of perceived stress (8–10). Perceived stress, which takes into account the perceived controllability, manageability, and personal impact of situations, has a strong and consistent association with concurrent and subsequent IBD symptoms (11). However, it is unclear whether stress induces increased levels of gut inflammation, or whether individuals with higher stress experience more intense symptoms independent of gut inflammation. Conversely, individuals with gut inflammation or a high symptom burden may have higher levels of perceived stress. Although the gold standard for determining the presence of active inflammation is through performance of endoscopy with histologic sampling (12), endoscopy is generally not immediately accessible for individuals who present to an outpatient clinic with troublesome GI symptoms. Standard biochemical tests such as serum C-reactive protein and erythrocyte sedimentation rate may suggest the presence of inflammation (13,14). However, these tests have limited sensitivity and specificity, and therefore many clinicians elect to treat symptoms without definitively confirming the presence of GI inflammation. A recent survey of gastroenterologists reported that the majority use symptomatic activity rather than the presence of intestinal inflammation (15) to guide therapy. Treating patient symptoms without confirming the presence and/or absence of inflammation may lead to unnecessary exposure to therapies that may have significant side effects as well as high costs. Recently, there has been increasing evidence that the level of fecal calprotectin (FCAL) is very sensitive and specific for the presence of intestinal inflammation in both CD and UC (16,17). FCAL is released from the breakdown of neutrophils in the intestines, and is not elevated in the absence of gut inflammation. It has been shown to be useful in differentiating symptomatic patients with IBD from those with functional GI disorders such as irritable bowel syndrome (18,19). FCAL levels are also well correlated with the degree of endoscopic and histological inflammation seen at endoscopy (19). However, the relationships among inflammatory activity, symptoms in IBD, and perceived stress are not well understood. Therefore, using a cross-sectional design and a population-based sample of individuals with IBD, we aimed to characterize the relationship among stress, symptom activity, and inflammatory activity, the latter determined through assessment of FCAL.
METHODS Sample
Participants were recruited from the University of Manitoba Research Registry (20). The Registry was developed using a The American Journal of GASTROENTEROLOGY
well-established administrative definition of IBD to identify all individuals in Manitoba with IBD. In 2000, and again in 2008, all individuals who met this administrative definition for IBD in the provincial single-insurer health database were contacted to invite them to be part of the Research Registry, with a response rate of ∼ 50%. Inclusion in the Registry is voluntary, and participants are not compensated for agreeing to be included. Individuals in the Registry agree to be contacted about various research initiatives, but are not obligated to participate. The demographic composition of the Research Registry is comparable to the broad population of individuals with IBD in Manitoba (21). At the time of recruitment, the Registry included 1,958 living individuals aged 250 μ g/g). Perceived stress was entered as a continuous variable, such that odds ratios (ORs) reflect the risk of increased IBD symptoms or intestinal inflammation associated with a 1-point change in level of perceived stress on the CPSS. We opted to employ the CPSS scores in this way to take advantage of the increased statistical power afforded by the use of a continuous scale. The P values of 2 units/day)
10 (3.8%)
10 (4.7%)
Prednisone
16 (6.1%)
10 (4.6%)
Biologic agent
37 (14.3%)
10 (4.7%)
3.9±1.5
4.2±1.6
93 (36.0%)
56 (26.2%)
5.6±4.1
4.3±3.6
132 (54.8%)
71 (39.0%)
20.9±8.7
19.4±9.4
Medication use in previous 3 months
Disease activity symptom scores MIBDI, mean (s.d.) MIBDI, proportion active (symptoms ≥ a few days every other week) n (%) Harvey Bradshaw/Powell Tuck Index, mean±s.d. Harvey Bradshaw/Powell Tuck, proportion active (≥5) n(%) Perceived stress Cohen Perceived Stress Score (mean±s.d.) IBD, inflammatory bowel disease; MIBDI, Manitoba IBD Index.
with lengthy disease (i.e., ≥20 years) had no current inflammation (63.7% vs. 48.0%, P=0.013). Considering the participants with UC, 34.5% overall were found to have elevated FCAL levels. Comparing those with normal and elevated FCAL, a considerably higher proportion of those with symptomatically active disease as measured by the MIBDI had an elevated FCAL level (42.7% vs. 17.3%, P0.2
N (%) ≥45
Mean±s.d.
127 (79.9%)
76 (74.5%)
>0.2
122 (85.9%)
61 (77.3%)
0.110
Female
113 (71.1%)
66 (64.7%)
>0.2
77 (54.4%)
47 (62.7%)
>0.2
7 (4.5%)
9 (9.0%)
0.153
5 (3.6%)
5 (6.8%)
>0.2
21 (13.4%)
16 (15.7%)
>0.2
5 (3.6%)
5 (6.8%)
>0.2
23.5±10.4
22.4±10.8
>0.2
22.2±11.8
18.0±9.8
0.009
Medication use in past 3 months Prednisone Biologic Duration of disease Mean years±s.d. N (%) with ≥20 years of duration
100 (63.7%)
49 (48.0%)
0.013
76 (54.7%)
27 (36.5%)
0.011
Current smoker
15 (9.5%)
15 (14.7%)
>0.2
9 (6.4%)
5 (6.7%)
>0.2
Alcohol use (≥2 units/day)
5 (3.1%)
5 (5.0%)
>0.2
5 (3.6%)
5 (6.7%)
>0.2
91 (57.6%)
55 (54.5%)
>0.2
10 (7.0%)
6 (8.1%)
>0.2
History of IBD-related surgery
FCAL, fecal calprotectin (μ g per g stool); HBI, Harvey Bradshaw Index; IBD, inflammatory bowel disease; IQR, interquartile range; MIBDI, Manitoba IBD Index; PTI, Powell Tuck Index. Proportions are reported as percentages of those with active inflammation or with low inflammation who reported a high level of symptoms. MIBDI: reports of symptoms occurring sometimes (a few days every other week), often (some days each week), or all the time (every day) were defined as actively symptomatic; higher scores on the MIBDI indicate less frequent symptoms. a N=258 in CD and n=214 in UC. Harvey Bradshaw Clinical Index for Crohn’s disease; scores of ≥5 were defined as actively symptomatic; higher scores indicated more frequent/severe symptoms. b N=241 for % comparisons and n=229 for mean comparisons. Powell Tuck Clinical Index for ulcerative colitis; scores of ≥5 were defined as actively symptomatic; higher scores indicate more frequent/severe symptoms. c N=182 for % comparisons and n=181 for mean comparisons. Perceived stress was assessed using Cohen’s Perceived Stress Scale; higher scores indicate higher perceived stress. d N=260 in CD and n=215 in UC.
Multivariate logistic regression was used to evaluate the relationship of several personal and disease characteristics with active symptomatic disease for respondents with CD and UC, and the results are shown in Tables 5 and 6 respectively. For the CD group, there was a greater likelihood of active symptomatic disease, based on the HBI, given a recent use of prednisone, a history of IBD surgery, current smoking, and higher concurrent perceived stress, controlling for the other factors. Active symptomatic disease based on the MIBDI was significantly more likely if a biologic medication had been used in recent months and with higher concurrent perceived stress. Intestinal inflammation (elevated FCAL) was not associated with active symptomatic disease, as measured by either the MIBDI or HBI. For the UC group, previous IBD surgery, © 2015 by the American College of Gastroenterology
current elevated FCAL levels, and current higher perceived stress levels were all significantly associated with active symptomatic disease across both measures (MIBDI and PTI), controlling for other factors. Finally, the relationship of these personal and disease characteristics with elevated fecal calprotectin was evaluated separately for CD and UC (Table 7) using logistic regression. Considering the results for the CD group, elevated FCAL level was not significantly associated with most of the variables in the analysis, including clinical factors such as age at diagnosis, history of GI surgery, current biologic use, current prednisone use, smoking status, or current level of perceived stress. Elevated FCAL was associated with shorter disease duration (i.e.,
