Medication Nonadherence and the Outcomes of
`
`Patients with Quiescent Ulcerative Colitis
`
`Sunanda Kane, MD, MSPH, Dezheng Huo, MS, James Aikens, PhD, Stephen Hanauer, MD
`
`PURPOSE: We conducted a prospective study to determine
`the effects of nonadherence with mesalamine among patients
`with quiescent ulcerative colitis.
`METHODS: We followed a cohort of 99 consecutive patients
`who had ulcerative colitis in remission for more than 6 months
`
`and who were taking maintenance mesalamine. Medication ad-
`herence rates were calculated based on pharmacy records and a
`validated formula. Nonadherence was defined as refilling less
`than 80% of prescribed medication. Patients were followed pro-
`spectively and evaluated either in clinic or via telephone at 6, 12,
`and 24 months. The primary outcome was clinical recurrence of
`ulcerative colitis. Proportional hazards models were used to ad-
`just for confounders.
`
`RESULTS: At 6 months, 12 patients (12%) had clinical recur-
`rence ofdisease symptoms, all ofwhom were nonadherent with
`medication. At 12 months, 19 of 86 patients had recurrent dis-
`ease, 13 (68%) of whom were nonadherent. Patients who were
`not adherent with medication had more than a fivefold greater
`risk of recurrence than adherent patients (hazard ratio = 5.5;
`95% confidence interval: 2.3 to 13; P < 0.001).
`CONCLUSION: Nonadherence with medication increases the I
`risk of clinical relapse among patients with quiescent ulcerative
`colitis. Future research should be directed at behavioral inter-
`
`ventions to improve adherence. Am J Med. 2003;114:39—43.
`©2003 by Excerpta Medica Inc.
`
`lcerative colitis is an idiopathic, chronic inflam-
`matory disease of the large intestine that is char-
`acterized by episodes of relapse and remission.
`As a chronic condition, therapy must continue on an in-
`definite basis to prevent relapse and to reduce the risk of
`long—term complications. Relapses are not predictable,
`although factors such as smoking cessation ( 1,2), psycho-
`logical stress (3), and possibly chronic u se of nonsteroidal
`anti—inflammatory drugs (4,5) may exacerbate symp-
`toms.
`
`Several large trials have demonstrated the efficacy of
`maintenance therapy for patients with quiescent ulcer-
`ative colitis (6-10). However, in contrast with other
`chronic illnesses (11-13), the potential effects of medica-
`tion adherence on recurrence of ulcerative colitis have
`
`not been evaluated. T/Ve previously reported that only
`about 40% of patients were adherent with maintenance
`therapies for ulcerative colitis (14). The aim of this study
`was to determine the effect of medication adherence on
`
`the clinical outcomes of patients with quiescent ulcer—
`ative colitis.
`
`From the Department of Medicine (SK, SH), and Department ofHealth
`Studies (DH), University of Chicago, Chicago, Illinois; and the Depart-
`ment of Family Medicine (IA), University of Michigan, Ann Arbor,
`Michigan.
`This research was funded by grants from Procter and Gamble Phar-
`maceuticals, Cincinnati, Ohio, and the David and Reva Logan Center
`for Gastrointestinal Research, Chicago, Illinois.
`Requests for reprints should be addressed to Sunanda Kane, MD,
`MSPH, Department of Medicine, University of Chicago, 5841 South
`Maryland Avenue, MC 4076, Chicago,
`lllinois 60637, or skane@
`medicine.bsd.uchicagoedu.
`
`©2003 by Excerpta Medica Inc.
`All rights reserved.
`
`METHODS
`
`Patients
`
`Patients followed at the University of Chicago Adult Gas-
`troenterology Outpatient Clinic were enrolled. Patients
`were recruited consecutively from May 1998 through Oc-
`tober l998, either during a clinic Visit or via telephone
`after a patient—initiated request for a medication refill.
`Eligibility criteria included a history of quiescent ulcer-
`ative colitis for at least the preceding 6 months and main—
`tenance treatment with mesalamine (Asacol, Procter and
`Gamble, Cincinnati, Ohio). Tl1e diagnosis of ulcerative
`colitis was verified by standard criteria ( 1 5). Patients with
`a hospitalization within the previous 12 months, who had
`used steroid or immunomodulatory therapy within the
`past 6 months, or who had undergone colitis surgery were
`excluded.
`
`Data Collection
`
`At the time of initial interview, demographic and clinical
`information was obtained from the patient and medical
`records. Demographic information included age, sex,
`marital status, education level, residential address, and
`
`insurance type. The telephone numbers for each phar-
`macy from which the patient had prescriptions filled (in-
`cluding mail order warehouses) were collected. We also
`recorded the length of the current remission, the length of
`time since the last use of steroids, the date of the last
`
`follow—up visit to a gastroenterologist, the date of the last
`colonoscopy, and niesalaniine dose and regin1en.Health—
`related quality of life was measured using the Short In-
`flammatory Bowel Disease Questionnaire (16). Use of
`other medications was also documented.
`
`In subsequent interviews, additional information was
`obtained about the recurrence of disease symptoms. The
`
`0002-9343/03/$—see front matter
`doi: l0.l0l6/S0002-9343(02)0l383—9
`
`39
`
`SAL|X00107800
`
`Dr. Falk Ex. 2016
`GeneriCo v. Dr. Falk IPR2016-00297
`Page 1
`
`

`
`Z\/Iedicatioiz Nonadhereizce and R€C14I‘I‘€I1C€
`
`of Ulcemtive Colitis/Kane et al
`
`date of the recurrence was recorded, along with symp-
`toms, including an increased number of bowel move-
`ments, rectal bleeding, urgency or cramps necessitating a
`change in treatment or a medical intervention, or changes
`in medications or dosages. Changes in medical history
`were also recorded (e.g., recent pregnancy or a new med-
`ical condition). ln addition, patients were asked to con-
`firm their prescription instructions, and were asked how
`much medication they were taking each day. Patients
`were given the opportunity to disclose reasons why they
`were 11ot taking their medications as prescribed.
`At the time of recruitment, and 6, 12, and 24 months
`after enrollment, a medication adherence rate was calcu-
`
`lated using a validated formula (17):
`
`Sum of days’ supply dispensed
`Sum of days in all refill intervals X 100
`
`The 6 months before each interview were used for pre-
`scription refill data, which were obtained by contacting
`pharmacies. The dates of medication refills were re-
`corded. Medication information was collected about me-
`
`salamine refills and other new prescriptions filled during
`that period. Nonadherence was defined as filling less than
`80% of prescribed medication (18).
`Before data collection by telephone, patients were
`asked to give verbal consent to proceed with the inter-
`view, as well as for calls to their pharmacy for medication
`information. The patients were told that the information
`provided was to help with better understanding of their
`disease. The investigator conducting the patient inter-
`views about clinical recurrence and quality of life was
`blinded to the patient’s adherence status and medication
`information. Patients who reported symptomatic recur-
`rence were asked additional questions about possible ex-
`acerbating factors.
`
`Smfiflkahhmhws
`The main outcome was clinical recurrence of disease
`
`(19); endoscopic data were not used in the scoring. Re-
`mission was therefore defined as one to three formed
`
`bowel movements per day, without any urgency, pain, or
`bleeding; recurrence was defined as four or more bowel
`movements per day associated with urgency, pain, or
`bleeding, or the presence of urgency, pain, or bleeding.
`The Wilcoxon rank—sum test was used to compare
`characteristics of patients with or without clinical recur-
`rence. A Cox proportional hazards model was used to
`identify variables that were associated with recurrence,
`including a time—dependent variable for medication ad-
`herence. The ‘/Vilcoxon signed—rank test and the Spear—
`man correlation coefficients were also calculated. Vari-
`
`ables with a P value <0.25 in univariate analyses were
`included in the model and added in a stepwise fashion. A
`Kaplan—Meier survival curve was constructed to compare
`outcomes stratified by adherence status. A P Value of 0.05
`(two—sided) was used to define statistical significance. All
`
`40
`
`January 2003
`
`THE AMERICAN )OURNAL or 1vI1:nrcrNE.® Volume 114
`
`Table 1. Characteristics 01°99 Patients with Ulcerative Colitis in
`Remission at Enrollment
`
`Characteristic (unit)
`
`Age (years)
`Male sex
`Married
`
`Disease duration (years)
`Length of remission (months)
`Quality—of—life score*
`Prescribed dose ofmesalamine ( g/d)
`Prescription coverage
`
`Number (96) or
`Median (Range)
`
`42 (18-79)
`52 (53)
`57 (58)
`8 (1-66)
`24 (6—360)
`62 (45-70)
`3.9 (1.2—4.8)
`60 (61)
`
`* Based on the Short lnflammatory Bowel Disease Questionnaire (16).
`
`analyses were performed using SAS software (SAS Insti-
`tute, Cary, North Carolina).
`The study was approved by the University of Chicago
`Institutional Review Board before patient enrollment.
`
`RESULTS
`
`Ninety—ni ne consecutive patients with ulcerative colitis in
`remission and who were being treated with mesalamine
`were recruited during an 8-month period (Table 1). Thir-
`ty—nine patients developed recurrent symptoms during
`follow—up. The annual incidence of relapse was about
`20°/0.
`
`By 6 months, 12 patients (12%) had clinical recurrence
`of their disease, all ofwhom were nonadherent with med-
`
`ication. The median percentage of prescribed me-
`salamine refilled was 51% (range, 0% to 76%), compared
`with 77% (range, 3% to 100%) for patients still in remis-
`sion (Figure 1, P < 0.001). Eighteen other patients were
`nonadherent but still in remission. By 12 months, 19
`(22%) of 86 patients had a recurrence of their disease, 13
`(68%) of whom were nonadherent. The median amount
`of mesalamine refilled was 74% (range, 0% to 98%) for
`those with recurrent disease versus 81% (range, 70% to
`100%) in those with quiescent disease (P = 0.09). By 24
`months, 1 patient died of causes unrelated to ulcerative
`colitis, and another underwent proctocolectomy for
`newly diagnosed dysplasia. Eight (12%) of the remaining
`66 patients had a recurrence, 6 of wh om were nonadher-
`ent.
`
`Of the 39 patients with recurrent disease, 5 required a
`short course of steroids for moderate disease activity, and
`2 required institution of rectal therapy; the remaining 32
`required increased doses of mesalamine. No patient re-
`quired hospitalization.
`The median amount of medication refilled in patients
`with recurrent disease was significantly lower than those
`with quiescent disease (50% Vs. 80%, P = 0.03). Thirty-
`two (82%) of the 39 patients with clinical recurrence at
`
`SAL|XOO107801
`
`Dr. Falk Ex. 2016
`GeneriCo v. Dr. Falk IPR2016-00297
`Page 2
`
`

`
`Zl/Iedicatimz Nonadhereizce and Recurrence of Ulcemtive Colitis/Kane et al
`
`Recurrence
`
`I No Recurrence
`
`PercentageofMedication
`
`RefilledinPrevious6Months
`
`12
`
`Foltow-up (Months)
`
`Figure 1. Medication consumption rates by disease activity.
`
`24 months were nonadherent, compared with 20 (34%)
`of the 59 patients who remained in remission (P = 0.01).
`The mean (i SD) daily amount of mesalamine refilled
`for those patients who had recurrent symptoms was 0.91
`: 0.35 g/d at 6 months, 1.22 i 0.8 g/d at 12 months, and
`2.04 i 1.1 g/d at 24 months. The 59 patients who were
`still in remission after 24 months were taking an average
`of 2.7 i 1.2 g/d of mesalamine. The difference between
`the two groups was statistically significant (P < 0.05) for
`each comparison. The daily dose of mesalamine that pa-
`tients reported taking ranged from 75% to 100%. The
`correlation between pharmacy data and patient disclo-
`sure was good (r = 0.83). V/Vhen asked why the medica-
`tions were not taken, 35 (50%) of the 70 who responded
`said they forgot, 21 (30%) said there were too many pills,
`and 14 (20%) did not think they needed so much medi-
`cine.
`
`ln univariate analyses, nonadherence, duration of dis-
`ease, and length of remission differed significantly be-
`tween those whose disease recurred and those who re-
`
`mained in remission. There were no differences in age,
`sex, marital status, education level, regimen, dose, quali-
`ty—of—life score, disease extent, or family history. In a mul-
`tivariate model, nonadherence, shorter duration of dis-
`
`ease, and shorter length of remission were associated with
`clinical recurrence (Table 2).
`Adherent patients had an 89% chance of maintaining
`remission compared with only 39% in those who were
`nonadherent (Figure 2, P = 0.001).
`
`DISCUSSION
`
`In this prospective study, we found that clinical recur-
`rence of ulcerative colitis was associated with nonadher-
`
`ence to prescribed mesalamine. A previous study, which
`followed patients for 48 weeks to identify risk factors for
`recurrent ulcerative colitis, did not find any association
`between medication adherence and clinical relapse (20).
`However, medication adherence was measured by asking
`patients and was greater than 95% for all participants.
`Direct patient inquiiy is inaccurate as patients often over-
`estimate medication use (21), as we observed in this
`study.
`We used pharmacy data to provide a more accurate
`estimate of medication consumption. \/Ve chose this un-
`obtrusive measure because more direct measures, such as
`
`tagged tracers, serum drug metabolite levels, direct obser-
`
`Table 2. Factors Associated with Disease Recurrence during 24 Months of Follow—up
`Hazard Ratio
`
`Variable
`
`(95% Confidence Interval)
`
`Nonadherence (yes vs. no)
`Length of remission (<12 vs. 212 months)
`Disease duration (<5 vs. 25 years)
`Positive family history
`
`5.5 (2.343)
`2.7 (1.2—5.8)
`2.4 (1.1—5.1)
`2.4 (1.0—5.8)
`
`January 2003
`
`THE AMERICAN JOURNAL or Mr.oic,iNE® Volume 114
`
`41
`
`SAL|XOO107802
`
`Dr. Falk Ex. 2016
`GeneriCo v. Dr. Falk IPR2016-00297
`Page 3
`
`

`
`Z\/Iedicatioiz Nonadhereizce and Recurrence of Ulcemtive Colitis/Kane at al
`
`1
`
`Adherent
`
`Nonadherent
`
`
`
`
`
`PercentageRemaininginRemission
`
`0.00
`
`Adherent (N)
`Nonadherent (N)
`
`O
`
`40
`59
`
`Time (Months)
`
`Figure 2. Nonadherence rates and clinical recurrence at 24 months.
`
`vation, and pill counts, are too intrusive, artificial, cum-
`bersome, and costly for routine clinical care. In addition,
`adherence is unlikely in patients who do not obtain reg-
`ular refills, whereas adherence is likely among those who
`do obtain regular refills, because most patients do not
`“stockpile” or discard medication. Estimates calculated
`from refill data have been shown to correlate with plasma
`drug level for phenytoin and diastolic blood pressure
`(21). We minimized bias by contacting every pharmacy
`that patients used, including mail order warehouses,
`We collected information on dates of refills and the
`
`onset of clinical recurrence to determine the temporal
`association between medication refills and disease recur-
`
`rence. However, not all patients who experienced a re-
`lapse were nonadherent, nor did all nonadherent patients
`experience a clinical recurrence. Nonadherence, a shorter
`disease duration, and a shorter remission before enroll-
`
`ment were independent predictors of clinical recurrence.
`Most patients experienced recurrence of symptoms
`within the first 12 months of disease remission, confirm-
`ing earlier results (22). Previous studies have also found
`similar rates of clinical relapse (23—25).
`Adherence with medication can be a problem even for
`short—term courses of therapy (26). In our study, the ma-
`jority of nonadherent patients cited forgetfiilness as the
`primary reason for nonadherence, whereas others re-
`ported that the number of pills was too many. Addressing
`these issues with patients in a nonconfrontational man-
`ner may lead to changed behaviors and an enhanced ad-
`herence rate (27).
`Even within the adherent group, however, the rate of
`medication refills decreased with time. ‘Ne found that an
`
`42
`
`January 2003
`
`THE AMERICAN JOURNAL or l\/lF.DICINE®
`
`Volume 114
`
`average dose of 2 g/d after 24 months of quiescent disease
`appeared to be inadequate to keep patients in remission,
`whereas a dose of at least 2.7 g appeared to be sufficient.
`However, this study was not designed to determine the
`appropriate dose to maintain remission.
`We studied patients with quiescent disease and who
`were only treated with one drug; these patients are per-
`haps most likely to be nonadherent. Patients who also use
`immunomodulators in addition to mesalamine are likely
`to have more active disease, which may improve adher-
`ence.
`
`In conclusion, we found that nonadherence with me-
`
`salamine is associated with clinical recurrence of quies-
`cent ulcerative colitis. Attention can now turn to deter-
`
`mining why patients are nonadherent, and to developing
`diseasespecific interventions to improve adherence.
`
`REFERENCES
`
`1. Silverstein MD, Lashner BA, Hanauer SB. Cigarette smoking and
`ulcerative colitis: a case—control study. Mayo Clirz Proc. 1994,69:
`425-429.
`
`. Eraga XE, Vergara M, Medina C, et al. Effects of smoking on the
`presentation and clinical course ofinflanirnatory bowel disease. Eur
`IGas£roenI1erolHq)at0l. 1997;9:683—687.
`. Levenstein S, Prantera C, Varvo V, et al. Stress and exacerbation in
`ulcerative colitis: a prospective study of patients enrolled in remis-
`sion. Am ]Ga5troemernl. 2000;95:1213—1220.
`. Evans IM, McMahon AD, Murray FE, et al. Non—steroidal anti-
`intlammatory drugs are associated with emergency admission to
`hospital for colitis due to inflammatory bowel disease. Gut. 1997;
`40:619-622.
`
`SAL|XOO107803
`
`Dr. Falk Ex. 2016
`GeneriCo v. Dr. Falk IPR2016-00297
`Page 4
`
`

`
`A/Iedicatimz Nonadliereace and Recurrence of Ulcerative Colitis/Kane et al
`
`. Felder IB, Korelitz BI, Rajapakse R, et al. Effects of nonsteroidal
`antiinflammatory drugs on inflammatory bowel disease: a case-
`control study. Am ]Ga5tr'oenterol. 2000;95:1949-1954.
`. The Mesalamine Study Group. An oral preparation of mesalamine
`as long—term maintenance therapy for ulcerative colitis. A random-
`ized, placebo—controlled trial. Ann Intern Med. 1996;124:204 -211.
`. Ardizzone S, Petrillo M. Molteni P, et al. Coated oral 5—aminosali—
`cylic acid (Claversal) is equivalent to sullasalazine for remission
`maintenance in ulcerative colitis: a double—blind study. I Clin Gas-
`troenterol. 1995;21:287-289.
`. Fockens P, Mulder CI, Tytgat GN, et al. Comparison ofthe efficacy
`and safety of 1.5 compared with 3.0 g oral slow—release mesalazine
`(Pentasa) in the maintenance treatment of ulcerative colitis. Eur I
`Gastroenterol Hepatol. 1995;7:1025-1030.
`. Green IR. Gibson IA, Kerr GD, et al. Maintenance of remission of
`ulcerative colitis: a comparison between balsalazide 3 g daily and
`mesalazine 1.2 g daily over 12 months. Aliment Pharmacol Ther.
`1998;1Z:1207-1216.
`. Miner P, Hanauer S, Robinson M, et al. Safety and efficacy of con-
`trolled—release mesalamine for maintenance of remission in ulcer-
`
`ative colitis. Dig Dis Sci. 1995;40:296 -304.
`. Sherman I, Hutson A, Baumstein S, Hendeles L. Telephoning the
`patient’s pharmacy to assess adherence with asthma medications by
`measuring refill rate for prescriptions. II-‘ediatr. 2000;136:532-536.
`. Liu H, Golin CE, Miller LG, et al. A comparison study of multiple
`measures of adherence to HIV protease inhibitors. Ann Intern Med.
`2001;134:968 -977.
`. Payne KA, Esmonde—White S. Observational studies of antihyper—
`tensive medication use and compliance: is drug choice a factor in
`treatment adherence? Curr Hyper/tens Rep. 2000;2:5l5-524.
`. Kane SV, Cohen RD, Aikens IE, Hanauer SB. Predictors of non-
`compliance with mesalamine in quiescent ulcerative colitis. Am ]
`Gastroenterol. 2001 ;96:2929-2932.
`. Miner P. Clinical features, course, laboratory findings, and complica-
`tions in ulcerative colitis. In: Kirsner IBM, ed. Inflammatory Bowel
`Disease. Baltimore, Maryland: W. B. Saunders; 2000:299-305.
`
`16. Irvine EI, Zhou Q, Thompson AK. Tl1e Short Inflammatory Bowel
`Disease Questionnaire: a quality of life instrument for community
`physicians managing inflammatory bowel disease. Am ] Gastroen—
`terol. l996;9l:l57l-1578.
`. Steiner IF, Koepsell TD, Fihn SD, Inui TS. A general method of
`compliance assessment using centralized pharmacy records. De-
`scription and validation. Med Care. 1988;26:814-823.
`. Sackett E, Snow IC. The magnitude of compliance and noncompli-
`ance. In: Haynes RB, Sackett DL. eds. Compliance in Health Care.
`Baltimore, Maryland: Iohns Hopkins Press; 1979:19-27.
`. Truelove S, Witts L. Cortisone in ulcerative colitis: final report on a
`therapeutic trial. BM]. 1955;2:1041-1044.
`. Riley S, Mani V, Goddman MI, Lucas S. Why do patients with
`ulcerative colitis relapse? Gut. 1990;3 1: 179 -183.
`. Steiner I, Prochazka AV. The assessment of refill compliance using
`pharmacy records: methods, Validity, and applications. I Clin Epi-
`demiol. 1997;50:105-116.
`. Bitton A, Peppercorn M, Antonioli DA, et al. Clinical, biological,
`and histological parameters as predictors of relapse in ulcerative
`colitis. Gastroenterology. 2001;120:13-20.
`. Langholz E, Munkholm P, Davidsen M, Binder V. Course of ulcer-
`ative colitis: analysis of changes in disease activity over years. Gas-
`troenterology. 1994;107:3-11.
`. Bresci G, Parisi G, Gambardella L, et al. Evaluation of clinical pat-
`terns in ulcerative colitis: a long—term follow—up. Int] Clin Pharma-
`col Res. 1997;17:17-22.
`. Misiewicz I, Lennard—Iones I11’, Connell AM, et al. Controlled trial
`ofsulphasalazine in maintenance therapy for ulcerative colitis. Lan—
`Cel. 1965;l:185-189.
`. Lee M, Kemp A, Canning A, et al. A randomized controlled trial of
`an enhanced patient compliance program for Helicobacter pylori
`therapy. Arch Intern bled. 1999;159:2312-2316.
`_ Levy R, Feld AD. Increasing patient adherence to gastroenterology
`treatment and prevention regimens. Am I Gastroenterol. 199994:
`1733-1742.
`
`Ianuary 2003
`
`THE AMERICAN JOURNAL or l\/lEDICINE® Volume 114
`
`43
`
`SAL|XOO107804
`
`Dr. Falk Ex. 2016
`GeneriCo v. Dr. Falk IPR2016-00297
`Page 5