Abdominal pain in quiescent inflammatory bowel disease.

After evaluating consecutive patients who had undergone an ileocolonoscopy (for restaging of IBD and/or to evaluate gastrointestinal symptoms) and completed concurrent validated pain-related surveys at our center, 122 individuals out of 275 total patients (44.4%; 65 females, 57 males) were identified as having quiescent disease (i.e., Mayo endoscopy sub-score or SES-CD of zero and histopathologic inflammatory score of zero) ( Table 1 ). There were 93 individuals with CD (32.7% ileal CD (L1), 14.2% colonic CD (L2), 53.1% ileocolonic CD (L3)). Twenty-six patients had UC (3.1% had proctitis (E1), 40.6% L-sided UC (E2), 56.3% pan-UC (E3)) and 3 individuals had an indeterminate version of IBD-associated colitis. Of this cohort, 74 patients (60.7%) were found to have QP-IBD. CD and UC exhibited similar rates of QP-IBD (60.2% vs. 62.5% respectively, p=0.890). Of the quiescent IBD patients with pain, 33 (44.6%) described mild pain, 28 (37.8%) described moderate pain and 13 (17.6%) described severe pain (based upon responses to the HBI). Quiescent IBD patients with abdominal pain were more likely to be female compared to their pain-free quiescent counterparts (63.3% vs. 40.7%, p<0.010). Age (43.8 years vs. 44.3 years, p=0.850), disease duration (12.9 years vs. 16.1 years, p=0.070), disease distribution (CD: L1 35.3% vs. 28.9%, L2 13.2% vs. 15.6%, L3 51.5% vs. 55.5% (p=0.770); UC: E1 5.0% vs. 0.0%, E2 45.0% vs. 28.6%, E3 50.0% vs. 71.4% (p=0.540)) and incidence of disease complications such as stricturing (61.4% vs. 61.1%, p=0.990) and fistulae (31.6% vs. 30.6%, p=0.990) were similar between the QP-IBD and QNP-IBD sub-cohorts respectively. Rates of prior IBD-related surgery were very similar (36.5% vs. 35.4%, p=0.990). Clinical scores for IBD-associated quality of life (SIBDQ) were significantly lower (38.2 vs. 58.3, p<0.001) and disease activity scores were higher (HBI: 8.3 vs. 3.6, p<0.001; SCCAI: 5.0 vs. 2.0, p<0.001) in QP-IBD patients. QP-IBD patients were also more likely to have significant coincident symptoms of anxiety and/or depression (71.6% vs. 25.0%, p<0.001). Inflammatory laboratory values (including WBC, platelet count, ESR and CRP) were similar between the cohorts. No other overt differences in demographics or clinical characteristics were identified. Current corticosteroid use was more common in the QP-IBD cohort when compared to the quiescent pain-free cohort (18.9% vs. 2.1%, p<0.010). Mesalamine (24.4% vs. 30.5%, p=0.450), immunomodulator (24.4% vs. 20.8%, p=0.830) and biologic (47.3% vs. 52.1%, p=0.710) use were each statistically similar between these cohorts. QP-IBD patients were much more likely to be active smokers (13.5% vs. 2.1%, p<0.050). They were also more likely to be actively using opiates (18.9% vs. 2.1%, p<0.010), antidepressants or anxiolytics (47.3% vs. 22.9%, p<0.010) and/or “other” pain medications (as defined above) (50.0% vs. 18.8%, p<0.001). Using a multivariable logistic regression analysis to evaluate the 122-patient quiescent IBD cohort described above, QP-IBD was independently associated with female gender (OR: 2.93 (1.11–7.68), p<0.050), anxious/depressed state (OR: 8.33 (3.01 – 23.05), p<0.001), and corticosteroid use (OR: 22.65 (2.39 – 214.70), p<0.010) ( Table 2 ). In order to identify characteristics that could help to differentiate QP-IBD patients, we also used endoscopic and pain-related survey data to identify a cohort of patients with active IBD who reported significant abdominal pain to compare them to. Using this approach, we identified 110 individuals who qualified as having active IBD with abdominal pain (AP-IBD). The AP-IBD cohort was composed of statistically similar proportions of CD, UC and IC (p=0.090) and the disease location distribution was very similar to that seen in the QP-IBD cohort (CD: 26.4% ileal (L1), 20.6% colonic (L2), 53.0% ileocolonic (L3); UC: 2.5% proctitis (E1), 32.5% left-sided (E2), 65.0% pan-colonic (E3)). AP-IBD patients demonstrated a similar gender distribution (53.6%, p=0.220), mean age (41.8 years, p=0.390) and mean disease duration (9.9 years, p=0.270) compared to the QP-IBD cohort. Prior incidence of IBD-related complications, including strictures (60.9%, p=0.990) and non-perianal fistulae (29.0%, p=0.850), and incidence of active EIM(s) (41.8%, p=0.880) were very similar between the cohorts. SIBDQ, HBI and SCCAI scores were also statistically similar between the cohorts ( Table 3 ). Several laboratory values were significantly different between these two groups. Specifically, AP-IBD patients exhibited a higher mean WBC (9.3 × 10 3 cells, p<0.010), platelet count (333.5 × 10 3 , p<0.001), CRP (2.4 mg/dL, p<0.050), and ESR (26.2 mm/hour, p<0.010) ( Table 3 ). Using the laboratory-associated upper limit of normal (ULN) for each of these tests, we also found that QP-IBD patients less frequently exhibited a significantly elevated WBC (10.6% vs. 31.3%, p<0.010), platelet count (8.5% vs. 35.0%, p<0.001), ESR (25.5% vs. 45.2%, p<0.050), and CRP (37.5% vs. 69.4%, p<0.010) when compared to the AP-IBD cohort. When we looked for a significant elevation in any one of these tests simultaneously, we found that AP-IBD patients had exceeded the ULN for at least one of these tests at least 76.8% of the time compared to 44.0% for the quiescent IBD group without pain (p<0.001).

We performed a retrospective analysis using information derived from the Intestinal Diseases Natural History Database at our institution between 1/1/2015 and 8/31/2018. This database includes clinical and research information related to the encounters of IBD patients receiving clinical management at a tertiary care referral hospital with a dedicated IBD center that cares for over 5000 patients with IBD. Written informed consent was obtained from each participant and all of this work was performed in accordance with the rules and regulations set forth by the local Institutional Review Board. In order to be included in this study, participants had to be adults (i.e., greater than 17 years of age) and have an established diagnosis of CD, UC or IBD colitis of indeterminate nature (IC), based upon standard clinical criteria routinely used to identify IBD [ 19 ]. They also needed to have undergone an ileocolonoscopy and completed simultaneous surveys on abdominal pain experience (including the Short Inflammatory Bowel Disease Questionnaire (SIBDQ) and Harvey-Bradshaw Index (HBI) or Short Colitis Activity Index SCCAI)). UC patients were excluded if they had undergone previous IBD-related colonic surgery. Quiescent disease was defined as a lack of any endoscopic evidence of luminal inflammation (i.e., a Mayo endoscopy sub-score for UC or Simple Endoscopic Score for Crohn’s (SES-CD) of zero, respectively). Presence of “significant inflammation” was defined as moderate to severe activity based upon findings during endoscopic evaluation (UC: using a Mayo endoscopy sub-score ranging from 0–3, with 0=no disease (“quiescent”), 1=mild disease, 2=moderate disease, and 3=severe disease; CD: using SES-CD scores of 7–15 for moderate disease and >15 for severe disease) and histopathological evaluation (hematoxylin and eosin (H&E) sections were blindly assigned an inflammatory score on a scale of 0–3 by a trained pathologist specializing in digestive disease). Of note, all individuals included in this study had undergone previous endoscopic and radiologic testing, and IBD location and type was based upon the findings of those tests. This information was used to make the most objective assessment of current gastrointestinal inflammatory status possible. To this end, all determinations about inflammatory state were based upon direct endoscopic evaluation of the mucosa and histologic assessment of tissue biopsies taken in areas of previously established disease activity. In order to increase the rigor of our patient pain experience assessment, abdominal pain ratings were based on responses to two separate questions answered on the day of ileocolonoscopy: 1) the fourth question in the SIBDQ (“How often over the past two weeks have you experienced abdominal pain?”, where patients respond using a frequency-based inverse Likert scale, with 1 representing pain “all of the time” and 7 representing pain “none of the time”), and 2) the second item from the HBI which assess current severity of abdominal pain and includes potential responses of 0 (“no abdominal pain”), 1 (“mild”), 2 (“moderate”) and 3 (“severe”). For the purposes of this study, clinically relevant abdominal pain was defined as a numeric rating of <6 on the SIBDQ pain score and a score of 1 or greater on the HBI pain score. Patients were asked to focus on abdominal pain experience unrelated to symptoms they may have experienced in the setting of their respective bowel preps. Of note, a smaller cohort (n=116) of our study patients with quiescent IBD had completed a Patient-Reported Outcome Measurement Information System (PROMIS) survey question (v.1.0) related to abdominal pain intensity (“What is your level of pain now?”). In order to further validate our approach to abdominal pain assessment, we compared the individual SIBDQ and HBI pain scores from each patient to the corresponding PROMIS score for each patient as derived from the question above. The SIBDQ pain score showed a strong negative correlation with the PROMIS score (r=−0.69, p<0.001) (as would be expected considering that the SIBDQ scores are based on an inverse Likert scale), while the HBI pain scores demonstrated a strong positive correlation with the PROMIS score (r=0.75, p<0.0001). Thus, we felt confident that we were obtaining an accurate assessment of current and recent patient abdominal pain experience. Additionally, no study participants had a history of or demonstrated objective evidence of extra-luminal causes of abdominal pain (including pancreatitis, symptomatic cholelithiasis, endometriosis) based upon the most recent laboratory and imaging studies or clinical evaluation. Age, gender, IBD duration, IBD extent/location (e.g., organ involvement, using the Montreal classification system), disease complications (including stricture, fistula, abscess and cancer development), extra intestinal manifestations (EIM) (including inflammatory arthritides, IBD-associated dermatopathies (including pyoderma gangrenosum), erythema nodosum, uveitis, episcleritis, and primary sclerosing cholangitis), endoscopic severity (as defined by the Mayo Index endoscopy sub-score), medication use (including antidepressant/anxiolytic, corticosteroid, mesalamine, immunomodulator (azathioprine, 6-mercaptopurine and/or methotrexate) and biologic therapy (infliximab, adalimumab, certolizumab, golimumab, vedolizumab, and/or ustekinumab)), surgical history, laboratory values (white blood cell count (WBC), sedimentation rate (ESR), C-reactive protein (CRP)), opiate (for management of IBD), “other” pain medications (acetaminophen, NSAIDs, dicyclomine and/or tricyclic agents), and tobacco use were abstracted. Presence of anxiety or depression symptoms was determined based upon responses to the Hospital Anxiety and Depression Scale (HADS) completed at the time of the clinical encounter, using anxiety or depression sub-scores of 8 or greater to indicate the clinically significant presence of each [ 20 ]. Data was extracted and analyzed using GraphPad Prism version 8 (San Diego, CA) or R 3.5.0 (R Core Team (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org .). Initially, demographic and clinical variables were compared using univariate analysis (e.g., student’s t- test, Chi-square test or Fisher’s exact test as appropriate) between two distinct cohorts: 1) IBD patients with quiescent disease with abdominal pain (hereafter referred to as “QP-IBD”), and 2) IBD patients with quiescent disease without abdominal pain (hereafter referred to as “QNP-IBD”). A multivariable logistic regression model was then performed incorporating each significant variable identified during the univariate analysis to examine the odds of developing hypersensitive IBD. Univariate analysis was also used to compare demographic and clinical variables between 1) QP-IBD patients and 2) IBD patients with active disease who simultaneously reported significant abdominal pain (referred to hereafter as “AP-IBD”) in order to evaluate for clinical factors that might be used to differentiate between these two groups. The primary endpoint for each of these analyses was QP-IBD (as defined above). Values listed represent means +/− standard error measurement (SEM), percentages or odds ratio (OR) with 95% confidence intervals (CI) unless indicated otherwise. Of note, data associated with two of these patient cohorts (QNP-IBD and AP-IBD) were included in a previous study published earlier this year that focused primarily on evaluation of a hypoalgesic IBD cohort [ 21 ].

We demonstrated that abdominal pain is very common in quiescent IBD, including both inactive CD and UC. The incidence of abdominal pain in our quiescent study cohort (including both CD and UC) was higher than that described in a previous study of abdominal pain in quiescent IBD [ 2 ]. Despite the lack of evidence for discernible differences in IBD type, location, complications (including EIM’s, strictures or fistulae), surgical history or objective measures of inflammatory activity, we also found that quiescent IBD patients with abdominal pain used corticosteroids, opiates, other pain medications and tobacco more frequently than the pain-free quiescent cohort. QP-IBD patients were also more likely to exhibit an anxious or depressed state and to use antidepressants or anxiolytics. Gender, anxious/depressed state, and corticosteroid use were each independently associated with abdominal pain in quiescent IBD. These findings are comparable to those reported in previous studies evaluating persistent symptoms in quiescent IBD [ 2 , 13 , 15 ]. Our study is the first to evaluate abdominal pain in consecutive quiescent IBD patients by pairing data from multiple pain surveys with simultaneous endoscopic assessment of disease activity. This approach provides a more objective assessment of disease activity than has been relied upon in prior investigations, which have primarily utilized clinical disease activity scores or physician global assessment [ 2 , 12 ]. It is also the first study to utilize multiple simultaneously administered abdominal pain surveys. Finally, this is the first study of its kind to evaluate consecutive IBD patients, which we believe provided a more dependable idea of pain incidence in this population. We believe that each of these study design elements enhanced the rigor and reliability of our study findings. Our investigation was also relatively unique because it included simultaneous assessment of particular key clinical factors that have the potential to influence patient pain experience, including prior surgical history, medication and drug use. Using this approach, we demonstrated that differences in pain reporting in the quiescent cohorts were not dependent on use of analgesics or other pain-modifying substances (e.g., marijuana). In fact, QP-IBD patients used pain-modifying agents more commonly than their pain-free counterparts. Taking these findings in coordination with those described above, they suggest that abdominal pain perception in quiescent IBD is likely influenced by factors other than disease type, severity or differences in medication or analgesic use. Another novel element of our study involved comparing quiescent and active IBD cohorts with abdominal pain. In so doing, we demonstrated that several inflammatory markers (including the WBC, platelet count, ESR and CRP) exhibit a higher mean value in AP-IBD patients and are pathologically elevated in this population more frequently than in QP-IBD patients. Unfortunately, almost a half of QP-IBD patients exhibited elevation in one or more of these tests while approximately one quarter of AP-IBD patients demonstrated no elevation. It is possible that these findings could have represented persistent, unrecognized IBD activity. We felt this was less likely, however, given that we were relying upon direct endoscopic and histologic assessment of the gastrointestinal areas known to be involved in each case. It is also possible that extra-intestinal factors could have contributed to this result (e.g., arthritides, dermatopathies), though that is unclear from this study. As such, these findings provide important insights into the limitations of commonly utilized laboratory tests to discern whether pain is being driven by active gastrointestinal inflammation. This reinforces the value of using other objective assessments of IBD activity, including endoscopic and radiologic testing, when patients report abdominal pain. Given the expense and frequent logistical challenges associated with completing these tests in a timely manner, though, there is still a tremendous need to develop more efficient methods to effectively screen for disease activity in IBD patients and to identify patients who may be at increased risk of developing abdominal pain. As indicated above, a large percentage of the quiescent patients in this study described abdominal pain at the time of and leading up to their endoscopic evaluation. It is not clear why such a sizable proportion of these individuals had pain and it is reasonable to question how clinically relevant this symptom was in some of these cases. In fact, 45% of these patients described having abdominal pain that was “mild” in severity. However, one third of all quiescent patients described moderate to severe abdominal pain. Importantly, QP-IBD patients exhibited a mean patient-dependent quality of life (SIBDQ) and clinical disease activity scores (HBI and SCCAI) that approximated those of active IBD patients with pain. All of these findings reinforce previous studies, which show that abdominal pain is common in this setting and has a significant, negative impact on patient disease perception. Our study also reinforces the concept that abdominal pain by itself can lead to increased use of consequential therapies, including opioid analgesics. A variety of factors have been implicated as drivers of persistent pain and associated gastrointestinal symptomatology in IBD. These include unrecognized luminal inflammation, underlying psychiatric conditions and exacerbating medication or substance use [ 2 , 12 , 13 , 22 , 23 ]. However, a large percentage of QP-IBD patients have no evidence for any of these issues [ 2 ]. Recent common practice has been to label these patients as having an overlap functional bowel disorder (e.g., IBS-IBD), particularly given the relatively common nature of these disorders [ 24 ]. However, this still leaves many questions about what actually drives pain (and other symptoms) in this condition. With this in mind, there are a variety of mechanistic theories relating to how abdominal pain develops and/or persists under these circumstances. There is evidence that UC patients may be hypersensitive to luminal stimuli, though there is some disagreement in the literature [ 25 , 26 ]. Several peripheral mediators of neuroendocrine function (including alterations in neurotrophic signaling factors, and intestinal serotonin signaling) as well as changes to the intestinal barrier function, have also been implicated in this regard [ 14 , 27 , 28 ]. Central modifiers of pain have been implicated as well, including alterations to hypothalamic-pituitary-adrenal axis function and differences in regional brain responses to anticipated or perceived visceral pain [ 29 , 30 ]. It is notable that QP-IBD patients had a higher rate of corticosteroid use. While this was found in a small minority of these patients, it is possible that at least some of the QP-IBD patients were just clearing their inflammation at the time of endoscopy and they were experiencing a post-inflammatory neuropathic phenomenon, potentially similar to what has been described in post-infectious irritable bowel syndrome [ 31 ]. Additionally, certain inflammatory cells or mediators that are less commonly evaluated for and/or identified in this context have demonstrated an emerging potential to induce pain anywhere along the course of the gastrointestinal tract (e.g., mast cell and eosinophil-related disorders) [ 32 , 33 ]. Mast cells may have a particularly compelling role in this regard [ 34 ]. While it is unlikely that there is a unifying cause of pain in this setting, these advancements have helped to shed new light on this issue. Nonetheless, the pathophysiology underlying pain in quiescent IBD remains unclear and larger and more mechanistic studies are required to advance our understanding of this issue. There are limitations to this study. First, it was undertaken in a single tertiary care referral center and it utilized data derived from a predominantly Caucasian population. As a result, these findings may not be relevant to all patients. Secondly, it is a relatively small study, evaluating even smaller sub-cohorts of CD and UC, potentially limiting our ability to identify otherwise significant clinical and/or demographic associations. Third, this was a retrospective study, thus some relevant clinical information may have been missed and there was the potential for recall or selection bias. This also made it more difficult to discern relative influences of different factors on abdominal pain in each case. For example, it is possible that increased corticosteroid use may have led to increased anxiety and/or depression in at least some patients, which in turn could have increased the likelihood of developing abdominal pain. Similarly, we did not have information on the timing of the initial diagnosis of relevant conditions (e.g., anxiety/depression) or the start of medications (e.g., corticosteroids) and this limited our ability to evaluate important potential clinical relationships. Fourth, we did not have consistent access to or results from other testing modalities that could have provided more comprehensive insight into luminal and extra-luminal intestinal inflammatory status, including stool-based markers of intestinal inflammation (e.g., fecal calprotectin) or abdominal/luminal imaging studies (e.g., CT/MR enterography). This is important, in part, due to the possibility of missing heretofore unrecognized deep small bowel or mesentery-based inflammation, which could be contributing to the symptoms [ 35 – 37 ]. Beyond this, we were also unable to definitively rule out some potential causes or modifiers of abdominal pain, including small bowel bacterial overgrowth, food intolerances (e.g., carbohydrate malabsorption) and intra-abdominal adhesions. Additionally, we excluded UC patients that had undergone surgery, limiting our ability to compare surgical rates as well as potential complication rates (e.g., involving strictures and/or prior pre-cancerous or cancerous lesions) among UC sub-cohorts. This study was also not designed to assess for potential cause and effect relationships and so we are unable to determine the exact cause(s) of pain these patients. Despite this, the results of this study are important because they reinforce that abdominal pain is common in quiescent IBD, regardless of sub-type, medication use or surgical history. We also identified risks associated with abdominal pain in this population, including inappropriate prescription of medications associated with major adverse effects, such as opiates and corticosteroids. These data highlight the importance of regularly utilizing objective markers of disease activity, particularly when considering use of the medications above. While laboratory markers of inflammation may be helpful, they are imperfect for differentiating patients with and without intestinal inflammation. Our study also highlights the need for developing more cost-effective and reliable methods for identifying patients at risk for developing abdominal pain in quiescent IBD. Finally, our findings reinforce the need to further investigate the drivers of abdominal pain in this setting. While many causes and modifiers have been invoked in these cases (including visceral hypersensitivity disorders, psychiatric comorbidities and medications such as NSAIDs and opioids), human visceral pain perception remains poorly understood and difficult to manage. To verify the findings of this study and to better address the mechanistic underpinnings of this pain, further larger scale human studies and laboratory-based models are required.

Individuals suffering from inflammatory bowel disease (IBD) frequently describe having abdominal pain [ 1 – 3 ]. Chronic abdominal pain is described in 70% or more of patients at the onset or during exacerbations of the disease and is one of the major reasons individuals with IBD seek out medical attention [ 1 , 3 ]. Abdominal pain has a significant impact on patient quality of life and it is estimated that this symptom costs the U.S. billions of dollars each year in health care expenditures and lost work hours [ 3 – 11 ]. The inflammation associated with IBD is considered to be the primary driver of pain, as pro-inflammatory cytokines and mediators sensitize extrinsic neurons that project to the gut [ 1 ]. However, inflammation alone does not explain pain in all IBD patients. Previous studies have suggested that up to 33% of individuals with ulcerative colitis (UC) and 57% of individuals with Crohn’s (CD) in clinical remission still may experience abdominal pain or other related symptoms [ 1 , 2 ]. However, the exact incidence and cause of abdominal pain in these individuals remains a mystery. There is evidence that many patients describing abdominal pain or irritable bowel syndrome (IBS) like symptoms actually exhibit persistent low-grade inflammation in the gut [ 2 , 12 – 14 ]. Several other studies have also demonstrated that many of these patients have coincident psychiatric disorders including anxiety, depression, and somatoform disorders that may modulate the visceral sensory and pain experience [ 2 , 6 , 15 ]. Each of these clinical scenarios listed is frequently associated with persistent pain in the IBD populations we have personally managed. Unfortunately, these same studies have frequently been limited by the lack of simultaneous evaluation of both inflammatory status and abdominal pain experience. There are also few studies that have specifically evaluated abdominal pain by itself rather than part of a constellation of “functional” or other gastrointestinal symptoms. Additionally, most investigations have utilized unreliable measures of disease inflammatory status, such as symptom-based clinical indices of disease activity (e.g., Harvey Bradshaw Index (HBI), Crohn’s Disease Activity Index (CDAI)) instead of the most effective assessment tools, such as endoscopy. Unfortunately, these approaches mirror a common clinical practice, where critical clinical decisions, such as starting corticosteroids, initiating potent analgesic medications, escalating IBD therapy, or hospitalizing patients, are made in response to patient-reported pain and other symptoms that may have no relation to the underlying status of the IBD [ 16 – 18 ]. In order to minimize these types of avoidable and potentially disastrous mistakes, and to better understand the scope of this issue, it is vitally important to have a more reliable evaluation of the patients who experience pain in quiescent IBD. We undertook this investigation to get a more accurate assessment for the incidence and patient characteristics associated with abdominal pain during quiescent IBD. We asked IBD patients to provide a thorough abdominal pain history while simultaneously undergoing ileocolonoscopy to properly assess IBD activity status. The primary aims of this study were to a) reassess how common abdominal pain was in a quiescent IBD population, b) identify risk factors associated with abdominal pain in the setting of quiescent IBD and c) elucidate characteristics that might differentiate these patients from patients with active disease experiencing pain.