Neuro-gastroenterology & Motility

Thirty-nine Italian referral centers for gastrointestinal disorders joined the laxative inadequate relief study (LIRS),12 a two-wave survey. In the first wave, LIRS 1, we enrolled 878 consecutive outpatients with chronic non-organic constipation from September through December 2011. In the second wave, LIRS 2, we enrolled 1325 outpatients in the same centers, from March 2012
and May 2013. Patients reporting at least two symptoms of FC according to Rome III criteria were evaluated by a gastroenterologist. 

The attending gastroenterologist recorded relevant clinical and demographic information in a standard data collection form.
All participating gastroenterologists were experienced in the
diagnosis and treatment of FGIDs and were instructed to exclude patients with secondary causes of constipation. Patients were classified in the FC or constipation-predominant IBS-C group, based on criteria listed in Table 1.
Patients willing to participate in the study and providing their informed consent were asked to fill in a self-administered survey.
We matched the medical data collection form and the selfadministered questionnaire for all study subjects.
 

 Outcomes (see also Appendix S1) We assessed patients’ healthrelated QoL with the SF-12,13 the severity of constipation with the patient assessment of constipation-symptoms (PAC-SYM),14 and
treatment satisfaction with the treatment satisfaction questionnaire for medication (TSQM-v2).15
The PAC-SYM is a 12-item questionnaire assessing the presence and magnitude of constipation-related symptoms worded as follows: (1) discomfort in your abdomen; (2) pain in your abdomen; (3) bloating in your abdomen; (4) stomach cramps; (5) painful bowel movements;  (6) rectal burning during or after a
bowel movement; (7) rectal tearing or bleeding after a bowel movement; (8) incomplete bowel movement, like you didn’t ‘finish’; (9) bowel movements that were too hard; (10) bowel movements that were too small; (11) straining or squeezing to try to pass bowel movements; (12) feeling like you had to pass a bowel movement you couldn’t. Ratings occur along a 5-point likert scale
(0 = absent; 4 = very severe). Items of the PAC-SYM questionnaire were identified through literature review and qualitative analysis of patient interviews (five focus groups of 8–10 patients
per group, separated by sex and by two levels of self-reported constipation severity). Of note, to maximize the relevance of the symptom assessment to patients, items were constructed from
the symptom description language used by focus-group patients. A recall period of 2 weeks was chosen for items to limit recall burden and to provide a relatively acute assessment. The instrument was designed for self-administration. Five pilot subjects completed the questionnaire to evaluate comprehensiveness and understandability. Furthermore, standard item reduction techniques were employed to select relevant items.14
 
The SF-12 is a 12-item questionnaire assessing patients’ perception of their own mental and physical health. Response to the questionnaire yield two composite scores (the Mental Composite,
MCS and the Physical Composite, PCS) ranging from 0 to 100, with higher scores indicating better QoL. The scoring system is norm based on the Italian general population (mean = 50; SD = 10).
The TSQM-v2 is an 11-item questionnaire assessing patients’ satisfaction for the efficacy, safety, and convenience of use of their treatment. Patients response yield three subscales (efficacy,
safety, and convenience) and an overall satisfaction rating. Scores occur on a 0–100 scale. Higher ratings indicate higher satisfaction with medication.
The LIRS II survey additionally inquired about therapy switching
in the 3 months priorto theinterview.We asked patients to rate their perceived overall clinical change after therapy switch with a bipolar Global Rating of Change Scale (GRC, 15 point, from
7 = extremely
deteriorated to +7 = extremely improved).16 As it has been previously suggested that the Minimal Clinically Important Change on a GRC scale is 2 points,16 we classified patients ratings of clinical
change asfollows:
7 :
5 = verydeteriorated;
4 :
2:deteriorated;

1 : 1 = unchanged; +2 : +4 = improved; +5 : +7 = very improved.
 

Definition of bloating Item 3 of the PAC-SYM was used to define
patients with (score = 1–4) and without bloating (score = 0).
 

Covariates The covariates included in the model were patients’ age, sex, employment, the diagnosis of IBS-C or FC based on ROME III criteria and the presence of abdominal pain and its
characteristics, BMI, smoking habit, daily intake of water greater than 1 L (yes/no), duration of lower gastrointestinal symptoms (years since diagnosis), number of pregnancies and difficult
deliveries, and the number of comorbidities and treatments (behavioral advice, bulking/osmotic, stimulant/herbal, enema/ suppository, prucalopride, any association, or no therapy). Comorbidities included all diseases treated or diagnosed in the past 12 months and previous abdominal and extra-abdominal surgical interventions occurred in the same time frame as well.
 

We defined questionnaire completers as all patients with no missing value in items relevant for questionnaire scoring. For each questionnaire item, the prevalence of missing values ranged  from 1% to 7%. We tested differences in socio-demographic and clinical factors (see Measures, Covariates) among completers and non-completers with chi-squared and Wilcoxon test where appropriate. As we found no statistically significant differences between completers and non-completers in both LIRS I and LIRS
II waves, we computed all questionnaire scores only for cases with complete answers.
 

Prevalence and magnitude of bloating 

The prevalence (PACSYM item 3 ≥ 1) and magnitude (PAC-SYM item 3 score) of
bloating have been computed in patients with FC and IBS-C. We evaluated unadjusted differences in the prevalence and magnitude of bloating with chi-squared and Wilcoxon test. We adopted logistic regression and generalized linear models to evaluate correlates of the prevalence and magnitude of bloating, respectively.
 

Bloating and patient-reported outcomes 

To test whether bloating was associated with poorer patient-reported outcomes (PRO), we
used Wilcoxon test to contrast QOL (SF12 summary scores),
Treatment Satisfaction (TSQMv2 scales), and GRC scores
between patients with and without bloating. We adopted generalized linear models to test the robustness of the results after adjustment for possible confounders and overlap-corrected PACSYM score (total score after deletion of the score reported in Item 3, the descriptor of bloating).
 

Sensitivity analysis 

Even though we advised all participating gastroenterologists to exclude any patient with alarm signs and
any organic disease potentially associated with constipation, we relied on each center clinical practice rather than providing a standardized diagnostic algorithm for the study. As a consequence, a bias in patient enrollment cannot be completely ruled out. To evaluate the potential impact of selection bias we
assessed differences in bloating prevalence, constipation characteristics (Rome III criteria), socio-demographic, and clinical characteristics between patients with and without organic diseases potentially associated with constipation (i.e., diabetes,\anal fissures, previous history of abdominal surgery, clinical
depression, chronic kidney disease, endometriosis, suspect heavy metals intoxication, cancers and neoplasms, hysterectomy,

Alzheimer’s disease or dementia, Parkinson’s disease, multiple
sclerosis, paraplegia or hemiparesis, diverticulosis, rectal prolapse; n = 654). As a further check, we included the interaction term between bloating and an indicator variable denoting a medical history
suggesting organic diseases potentially associated with constipation in all models evaluating the association between bloating and PRO (SF-12, TSQMv2, overlap-corrected PAC-SYM). A statistically significant interaction term would suggest possible selection bias. This procedure allows to estimate both the magnitude and the direction of possible selection bias. We report the differencein-difference of the adjusted means calculated as follows:

 

Patients’ characteristics are described in Table 2. LIRS 1 and LIRS 2 were substantially similar relative to key patients’ characteristics. About 55% (n = 1212) received dietary prescriptions alone or in associations with other treatments. Additionally, 512 patients (23.2%) received fiber supplementation alone or in associations with other treatments. There was no difference across bloating groups in the proportion of patients receiving dietary prescriptions or fiber supplementation (data not shown). Most patients received associations of constipation-related treatments in the month prior to the interview (i.e., schemes including a combination of more than one drug with or without dietary recommendation). Overall patients receiving dietary recommendations or fiber supplementation alone or in association with other treatments were 208 (12.7%) and there was no difference in the proportion of fiber supplementation prescription across bloating strata. During LIRS I, fielding prucalopride
was not available in Italy. During LIRS II, 151 (11.4%) patients received a prescription for prucalopride alone and 167 (12.6%) patients in association with other treatments. Most common prucalopride associations were dietary (16.4%), bulking/osmotic (14.8%), enema/ suppository (13.5%), and stimulant/herbal (7.9%) treatments. Prescription of prucalopride was a switch from previous treatments in the majority of patients (94%).
Mean overlap-corrected PAC-SYM score (i.e., total score after deletion of the score reported in Item 3, the PAC-SYM descriptor of bloating) was 1.46  0.77.
 
 

The prevalence of bloating was 91.6% (n = 1970). Patients with IBS-C reported slightly more bloating
than patients with FC (95.9%, n = 347 vs 90.8%, n = 1623 in IBS-C and FC, respectively; p < 0.01). After
adjustment for possible confounders (covariates in the methods section), IBS-C (OR = 2.80, p < 0.01), women (OR = 2.36, p < 0.01), obese subjects (BMI >30) (OR = 3.51, p < 0.01) and patients with more comorbidities (>3: OR = 2.04; 1–3: OR = 1.96; 0: ref.; p < 0.01) were more likely to report any bloating. The magnitude of bloating (PAC-SYM item 3) was greater in patients with IBS-C compared to FC (2.69  0.93 vs 2.38  0.98, p < 0.01, respectively). After adjustment for possible confounders, patients with a diagnosis of IBS-C (b = 0.32, p < 0.01), women (b = 0.56, p < 0.01), those with a paid job (b = 0.15, p < 0.01), older subjects (less than 33 years: reference; 33–67 years: b =
0,14; older than 67 years:
0.34; p < 0,01), those currently smoking (b = 0.14, p < 0.01), and those with more comorbidities (0 medical conditions: b =
0.59; 1–3 medical conditions: b =
0.20; >3 medical conditions:
reference; p < 0.01) reported higher bloating magnitude. A diagnosis or treatment for depression/anxiety in the past 12 months (b = 0.12, p < 0.053) was marginally associated with bloating magnitude
 

Patients with bloating reported lower QOL scores independently of the diagnosis of FC or IBS-C (Fig. 1,
Panel A). Additionally, we observed that the relationship between bloating and HRQOL was not different in patients with FC or IBS-C (p for interaction >0.1). This pattern of association was robust to adjustment for possible confounders (Table 3). The association between bloating and SF-12 physical composite (Bloating: b =
2.34, p < 0.01) was independent of other constipation-related symptoms (overlap-corrected PAC-SYM scores: b =
3.83, p < 0.01). On the contrary the association between bloating and SF-12 mental composite (Bloating: b =
0.64, p = 0.46) was not independent of other constipation-related symptoms (overlap-corrected PAC-SYM score: b =
4.69, p < 0.01).
 
 

Among patients without bloating, 78% (n = 112) reported to be partially or completely satisfied with
the effectiveness, 96% (n = 136) with the safety and 91% (n = 127) with the convenience of administration of their treatment. Patients with bloating reported lower satisfaction on all dimensions (59%, 87%, and 82% respectively, p < 0.01). 

Additionally, patients with bloating reported lower treatment satisfaction scores independent of the diagnosis of FC or IBS-C (Fig. 1, Panel B). Bloating was associated with treatment satisfaction independent of possible confounders (Table 3). After adjustment for overlap-corrected PACSYM scores, bloating was still associated with TSQM
Effectiveness (b =
4.79, p < 0.01) and TSQM Convenience (b =
4.63, p = 0.02), while the association
with TSQM Safety was strongly attenuated and lost statistical significance (b =
2.82, p = 0.23).
 
 

A minority (n = 413, 30%) of LIRS II patients initiated a new therapy in the 3 months prior to the interview.
Most patients (66%) reported subjective clinical improvement after switch or initiation of therapy,
27% reported no clinical change and only 7% reported worsening symptoms. Bloating was strongly associated with patients’ rating of change after therapy switch independent of treatment and adjustment for possible confounders (Table 4).
 

Patients with a medical history of conditions possibly associated with secondary constipation were older (56.7 vs 47.4, p = 0.02) and more likely to report a feeling of incomplete evacuation in at least 25% of defecations (78% vs 71%, p < 0.05), a sensation
of anorectal obstruction/blockage for at least 25% of defecations (48% vs 36%, p < 0.05), the need to adopt manual maneuvers to facilitate at least 25% of defecations (22% vs 30%, p < 0.05). We observed no differences in bloating prevalence, IBS prevalence, gender, fluid intake >1 L/day, smoking habit, treatment, time since constipation onset across patients with/without previous history of organic diseases potentially associated with constipation (data not
shown). The relationship between bloating and health-related QoL (SF-12) was not moderated by the
inclusion/exclusion of patients with medical history of conditions possibly associated with secondary constipation (SF-12, PCS: p for interaction = 0.79; SF-12, MCS: p for interaction = 0.20). On the contrary, moderation analysis revealed that the association between bloating, treatment satisfaction, and other constipation-related symptoms was weaker in patients with a medical history of conditions possibly associated with secondary constipation. As such, the inclusion of patients with organic constipation in the analysis may lead to underestimating the association between patients’ report of bloating and constipation severity (overlap-corrected PAC-SYM: Dp-s = 0.49, p < 0.01), satisfaction with treatment effectiveness (TSQMv2- Effectiveness: D p-s = 12.38, p < 0.01), and satisfaction with treatment convenience of administration (TSQMv2-Convenience: Dp-s = 14.39, p < 0.01).
 
 
 

Dr Neri has received an unrestricted research grant by DoxaPharma SRL to support the production of this manuscript.
 

LN participated in study concept and design, data analysis, interpretation of results, manuscript drafting, and approved the final version of the manuscript; PI participated in study concept and design, data collection, interpretation of results, provided critical revisions to manuscript drafts, and approved the final version of the manuscript. All members of the LIRS study group participated in study concept and design, data collection, provided critical revisions to manuscript drafts, and approved the final version of the manuscript.