Physician Global Assessments or Blood tests do not Predict Mucosal Disease Activity in Ulcerative Colitis

BACKGROUND: Mucosal healing has been proposed as the therapeutic end point in the treatment of patients with ulcerative colitis (UC).


METHODS
Colonoscopy reports of all patients known to have UC from July 2008 to November 2012 from a single gastroenterology practice at the Health Sciences Centre, Winnipeg, Manitoba, a tertiary care referral centre for inflammatory bowel disease (IBD), were retrospectively reviewed.Patients included individuals with UC having bloodwork (complete blood count, CRP, ferritin, albumin) within one week of visit and undergoing colonoscopy after clinic assessment (within one month).Cut-off values for bloodwork in the analysis included Hg levels <120 g/L for females and <140 g/L for males, plt count >4.0×10 9 /L, albumin <33g/L, ferritin <20 µg/mL and CRP >8 mg/L, which represented abnormal values at the authors' institution.No deviations from standard care were taken (ie, patients being expedited to endoscopy).Patients presenting through the emergency room, those with colonoscopies performed outside the authors' institution, or whose colonoscopies and clinical assessments were undertaken more than one month apart were excluded.The PGA was used to determine disease activity in patients before colonoscopy and was performed at time of clinic visit.Currently, no validated PGA exists; therefore, an assessment tool based on routine questions gastroenterologists use during interviews to gauge disease activity was created.To determine a global assessment, patients were queried at each clinic visit as to the number of bowel movements, presence or absence of abdominal pain, presence of blood with defecation and objective weight loss (as determined by the clinic scale).These criteria used in the PGA are documented in all patient encounters at the authors' centre.The assessment was rated as: 1. Remission: no abdominal pain, ≤2 bowel movements/day, absence of blood with defecation, and stable/no weight loss.2. Mildly active disease: mild to moderate abdominal pain, ≤4 bowel movements/day, occasional blood with defecation and objective weight loss.
3. Active symptoms: moderate to severe abdominal pain, >4 bowel movements/day, bloody defecation with most bowel movements and objective weight loss.Patient disease activity was categorized based on having at least two of the four symptoms identified with each category.If they had only one symptom that moved them to a higher, more active category, it would not have changed their categorization.For example, an individual with one mild symptom and three symptoms that would have been considered to be in remission were categorized as remission.If at least two mild symptoms were present and zero or one active symptom, then they were considered to be mild.If there were at least two active symptoms they were categorized as active.For example, if a patient had mild abdominal pain (mild symptom), six bowel movements/day (active symptom), minimal blood with defecation (mild symptom) and no weight loss, they would be categorized as having 'mildly active' disease.Patients were identified through an electronic database of all persons presenting to the IBD clinic and their charts were systematically reviewed.Demographic data including age, sex and blood tests (serum Hg, plt count, CRP, ferritin and albumin level) were collected.Patients were grouped as active (active symptoms) or mild to inactive (remission, mild) for the PGA.A total of 154 patients met the inclusion criteria for the study.An adapted form of the Ulcerative Colitis Endoscopic Index of Severity (UCEIS), a validated scoring system to rate endoscopic disease activity in UC (Table 1), was used because the UCEIS was not available during the study period.Evaluating the most diseased area on endoscopy, the following parameters were used and categorized patients into: 1. Remission: complete MH that could include 'footprints' of past disease such as pseudopolyps or white scars.2. Mild disease: vascular blush or loss of vascular pattern, minimal exudates or friability.3. Moderate disease: friability, granularity, scattered erosions and ulcers.4. Severe disease: contiguous or deep ulcers and frank bleeding.
Similar to the PGA, the authors use a standard reporting format with all UC patients and include these elements in their colonoscopy reports in addition to pictures documenting diseased areas.Because the authors evaluated the most diseased areas, every endoscopic report included information regarding the most diseased area; the reports at their centre included the necessary variables to report a UCEIS.If there was insufficient information on the endoscopy report, the subject was excluded.The primary end point was to assess how well the PGA could assess MH as defined by features on endoscopy.Secondary end points included a correlation of blood tests (Hg, plt count, CRP,

RESULTS
A total of 154 patients with UC were identified.Males comprised 53% (n=82) of the total population.The mean (± SD) age of patients was 44±15.7 years.The mean Hg level was 139 g/L, mean platelet level was 296×10 9 /L, mean ferritin level was 102 µg/L, mean CRP level was 10 mg/L and mean albumin level was 40 g/L (Table 2).The average disease duration was 15 years (range two to 42 years), with 82% having pancolitis and 18% having left-sided colitis (including proctosigmoiditis) as the predominant disease extent at time of diagnosis.Of the colonoscopies, 76% were performed as routine dysplasia surveillance and 24% were performed due to patient experiencing symptoms such as abdominal pain, diarrhea or bright red blood per rectum.In terms of medications used, 46% were using 5-aminosalicylic acid products with varying doses as their predominant maintenance regimen, 22% were not on any medications, 13% were on immunomodulator therapy (azathioprine or 6-mercaptopurine), 8% were on anti-tumour necrosis factor (Remicade [Janssen Inc, USA] or Humira [Abbott Laboratories, USA]), 2% were on methotrexate and 9% were on combination therapy (Table 3).Using endoscopy as the 'gold standard' for assessing UC activity (moderate-severe), abnormalities in laboratory parameters were highly specific (Hg 88%; albumin 97%, ferritin 91%, plt count 89%, CRP 72%) but not sensitive (Hg 34%; albumin 7%, ferritin 23%, plt count 16%, CRP 28%) for identifying individuals with at least moderately active endoscopy.The PGA was comparably sensitive (30%) and specific (97%) as laboratory parameters but with higher positive (83%) and negative (74%) predictive values (Table 4).Also conducted was a subgroup analysis of patients experiencing symptoms, either 'mild' or 'active' according to the PGA, excluding those considered to be in remission, and found similarly high specificities for laboratory parameters (Hg 91%, albumin 92%, ferritin 72%, plt count 84%, CRP 86%) but still low sensitivities (Hg 45%, albumin 12%, ferritin 24%, plt count 23%, CRP 36%).The PGA also had comparable sensitivity (48%) and specificity (86%) but with higher positive (83%) and lower negative (54%) predictive values than the when analyzing all persons including those in remission (Table 5).

DISCUSSION
Traditionally, initiation and escalation of therapy in UC has been based on severity of symptoms with the goal of alleviating those symptoms using a variety of medications (5-aminosalicylic acid, steroids, immunomodulators and biologics).However, treating symptoms alone may not be sufficient to achieve optimal long-term outcomes (eg, fewer complications and hospitalizations, and better quality of life), and targeting mucosal inflammation and associated tissue damage may be equally important.MH has been proposed as an optimal goal of treatment because it can correlate with hospitalizations and colectomy rates (7)(8)(9)(10).The exact definition of MH is unclear and has not been validated; however, the International Organization for the Study of Inflammatory Bowel Disease has proposed defining MH in UC as the absence of friability, blood, erosions and ulcers in all visualized segments of gut mucosa (12).Whether MH should represent the complete absence of any characteristic endoscopic lesions or simply marked improvement in the severity of previously noted lesions is an issue that is evolving.Currently, no guidelines have been developed to determine the optimal timing for follow-up endoscopy to identify MH (13).Endoscopic evaluation can be time consuming and costly, while incurring small but real risks for adverse events.Froslie et al (7) documented the important role of MH in monitoring treatment effectiveness and long-term disease outcome during a five-year follow-up period.In their study, 50% of patients with UC had confirmed MH after one year and had a significantly lower risk of future colectomy than patients without MH (P=0.02).Several other long-term benefits of MH have been identified, including decreased need for surgery and hospitalization, lower steroid use, decreased risk of colorectal cancer and higher remission rates (5-10).
The PGA or, less formally, a clinician's impression of the patient's disease status, is typically used to guide therapeutic decisions.Many activity indexes have been used indirectly to assess clinical status.Common indexes, such as Truelove and Witts, Powell-Tuck, Ulcerative Colitis Disease Activity Index and Mayo Disease Activity Index, use a combination of symptoms, signs and sigmoidoscopy to assess clinical disease; however, none of these scales have been validated specifically for MH (14)(15)(16)(17)(18)(19)(20).Attempts are now underway to develop a validated endoscopic score that may be universally adopted such as the UCEIS developed by Travis et al (11).In our study, we found the PGA to correlate poorly with active endoscopic disease; however, it was marginally better than the blood tests routinely ordered by gastroenterologists to estimate disease activity status.When the PGA or any of the blood tests are abnormal, there was a high likelihood the endoscopic findings were at least moderately active.When the PGA or blood tests were normal, however, they missed many subjects who had active endoscopic disease.
Noninvasive quantitative indexes have been developed in UC based on symptoms and blood tests (Hg, albumin and erythrocyte sedimentation rate) that have correlated well with the Truelove and Witts classification (mild, moderate, severe) (21).However, a previous study involving 82 patients showed that CRP is only raised 50% of the time in active UC but does aid in differentiating IBD from chronic abdominal pain (22,23).In addition, blood tests such as Hg, plt  Our analysis highlights a few key points.First, physicians are not adequately able to predict endoscopic disease with reasonable sensitivity for UC based on their PGA or the 'usual' blood tests that are ordered.However, abnormal PGA or laboratory parameters do signal active endoscopic disease.The main limitation of our study was that it was retrospective in design, with all of the biases of retrospectively grading endoscopy scores and PGA.That a single physician performed all of the endoscopy and had a uniform pattern of assessment, including clinical details during the assessment, was a strength of the study; however, a single physician's practice limits the external validity of the results.Additionally, most patients underwent routine screening endoscopy (76%) and were not being assessed for acutely active symptoms.

DISCLOSURES:
The authors have no financial disclosures or conflicts of interest to declare.

CONCLUSION
A combination of serum markers (Hg, plt count, and ferritin, albumin and CRP levels), and clinical symptoms (PGA) in a comprehensive activity index may be a promising, noninvasive and, possibly, cost-effective approach to evaluate patients with UC; however, individually cannot replace endoscopy for assessing MH.Nonetheless, when patients experienced active symptoms and abnormal serum markers, they were highly likely to have abnormal endoscopy in our study.However, inactive symptoms or normal laboratory values did not preclude having active endoscopic disease.It is possible that other blood markers or fecal markers of inflammation may add to the PGA to serve as better surrogate markers of MH in future prospective studies involving adult patients with UC.

TAble 1 Ulcerative Colitis endoscopic Index of Severity (UCeIS) Descriptor (score of most severe lesions) Scale anchor points Definition
Deeper excavated defects in the mucosa, with a slightly raised edge UCEIS = sum of scores, which accounts for 92% of the variance between observers in the overall assessment of endoscopic severity ferritin and albumin) with endoscopic features.Because the present analysis was a retrospective observational study, results are presented with descriptive statistics.Mean and SD are used to report continuous variables following a normal distribution and median (range) are used to report non-normal continuous variables.

TAble 3 Ulcerative colitis characteristics
, ferritin levels and CRP can serve as markers to assess disease activity, but in the case of a hospitalized or sick patient, low or high laboratory markers can reflect an acute or chronic inflammatory process such as infection, malabsorption, trauma, cancer or the disease itself giving false positive/negative results.Unfortunately, none of these disease activity indexes, invasive or noninvasive, endoscopic or histological, has been formally well validated in terms of reflecting the evolution of the disease in the long term. counts