Secondary Logo

Journal Logo

ONLINE ARTICLE: Original Article

Comparative Efficacy of Drugs for the Treatment of Chronic Constipation

Quantitative Information for Medication Guidelines

Zhang, Yi MS; Yin, Fang PhD; Xu, Ling PhD; Li, Yun-fei PhD; Chen, Jun-chao PhD; Liu, Hong-xia PhD; Zheng, Qing-shan PhD; Li, Lu-jin PhD

Author Information
doi: 10.1097/MCG.0000000000001303

Abstract

Chronic constipation is one of the most common and frequently diagnosed gastrointestinal disorders1,2; its prevalence is ∼12% to 19% among the general population.3 Every year, millions of outpatient clinic visits in the United States, as well as 10 million prescriptions in the United Kingdom, are due to constipation.4 The treatment of this condition consumes a large number of medical resources and money. Making changes in the patient’s diet and lifestyle is the preferred method for relieving constipation.5–7 However, for patients who fail to obtain symptomatic relief of constipation through diet and lifestyle modifications, pharmacotherapy is usually required.8,9

Currently, the drugs used to treat constipation fall into the following categories: osmotic laxatives, stimulant laxatives, chloride channel activators, guanylate cyclase-C agonists, serotonin 5-hydroxytryptamine 4 receptor agonists, and ileal bile acid transporter inhibitors.10–13 However, relevant quantitative information is scarce in the current constipation medication guidelines, and these guidelines do not accurately reflect differences in the therapeutic efficacies of the various drugs.

A published meta-analysis indirectly compared the efficacy of 8 medications for the treatment of constipation. However, uncorrected heterogeneities between trials, such as different treatment durations (4 to 24 wk) and different demographic characteristics, made the conclusions ambiguous.14 Through the analysis of preliminary data, we found that the increased frequency of SBMs and CSBMs associated with some medications used to treat constipation exhibited a declining trend after a period of treatment. Unfortunately, previous studies have not analyzed this trend.

Model-based meta-analysis (MBMA) is a technique that involves pooling the results of multiple studies to gain a better overall understanding with regard to the pharmacodynamic profiles of different agents.15,16 Compared with conventional methods of meta-analysis, MBMA may be used to describe the dose-response characteristics, determine the efficacy of drugs over a period of time, and explore factors that influence these characteristics.17–19 In this study, we used MBMA to establish a pharmacodynamic model to quantitatively describe the time course of each drug used for chronic constipation. By creating this model, we aimed to identify relevant impacting factors and to use pharmacodynamic parameters to directly reflect the changes in the frequency of bowel movements associated with each drug, especially the characteristics as regards the loss of efficacy. The results will provide quantitative information to improve medication guidelines for the treatment of constipation.

Methods

Search Strategy

A comprehensive search was conducted in the PubMed, EMBASE, and Cochrane Library databases from the date of inception to January 2, 2018. Search keywords included the names of drugs and therapeutic indications. We limited our search to published studies written in English and clinical trials. The full details in regard to our search strategy are described in the Supplemental Digital Content 1 (http://links.lww.com/JCG/A540).

Only studies that met the following criteria were included in the analysis: (i) randomized, placebo-controlled trials, (ii) studies involving participants suffering from chronic constipation, and (iii) manuscripts that reported the frequencies of spontaneous bowel movements (SBMs) or complete spontaneous bowel movements (CSBMs) at baseline and after drug treatment.

The exclusion criteria were as follows: (i) studies involving pregnant women, puerperae, lactating women, children, and adolescents (age below 18 y); or (ii) studies involving participants diagnosed with irritable bowel syndrome, secondary constipation, or constipation caused by other diseases.

Data Extraction

Microsoft Excel software (version 2016) was used as a data entry platform. The following information was extracted from the included studies: study identifiers (authors, published year, and clinical trial registration number), trial design (treatment drugs, dosage, sample size, and duration of treatment), demographic characteristics [age, body mass index (BMI), percentage of female participants, and the baseline frequencies of SBMs or CSBMs], and clinical outcomes (frequencies of SBMs or CSBMs at each observation point).

Data with regard to the frequencies of SBMs or CSBMs were presented in a graph. Engauge Digitizer (version 4.1, 2002, by Mark Mitchell), a digitizing software, was used to extract data. All the above data were independently extracted by 2 researchers (Y.Z. and F.Y.), and data extraction was checked by a third person (L.X.). The error rate for data extraction did not exceed 2% while reading the graph. If the error rate exceeded 2%, the graph was reread, and the mean values were used for the final results.

Risk of Bias Assessment

The methodological quality of the included studies was assessed independently by 2 investigators (Y.Z. and F.Y.) using the risk of bias tool. This tool evaluates the risk of bias for randomized controlled trials using the following 7 domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other sources of bias. Each item was graded as providing a low, high, or unclear risk. Conflicting opinions were adjudicated by the third investigator (Y.-f.L.).

Model Building

In this study, the changes in SBM and CSBM frequencies from the baseline frequencies were used as the efficacy indices. To eliminate the effect of dietary and lifestyle heterogeneity on the efficacy of the therapeutic drug, we established models based on the true drug effect, that is, the difference in efficacy between the effect of a drug and its corresponding placebo. We found that the change in the frequencies of SBMs and CSBMs from the baseline values was greatest at the beginning of drug treatment. However, these frequencies gradually decreased over time, thus yielding losses of efficacy. This pharmacodynamic characteristic can be described using the following equations:20

In Equation 1, Ei,j is the observed changes in frequency from baseline SBM and CSBM frequencies at the timepoint j in the ith drug group. Emax,i represents the theoretical maximal frequency change from baseline SBM and CSBM frequencies in the ith drug group. τi is the loss of efficacy rate in the ith drug group; Time represents the observation timepoint. εi,j is the residual error in the ith drug group at the timepoint j and was weighted by the inverse of the square root of the sample size (Ni,j). εi,j was assumed to be normally distributed with a mean of 0 and variance of σ2. In Equations 2 and 3, Emax,typical and τtypical are typical values for the pharmacodynamic parameters Emax and τ of the overall drug group, respectively. η1,i and η2,i represent the intergroup variability for Emax and τ, respectively. Emax and τ, are assumed to be normally distributed, with means of 0 and variances of ω12 and ω22.

A nonlinear mixed-effects model (NONMEM) was used to fit the effect data. Pharmacodynamic parameters (ie, Emax,i and τi) for each drug group were obtained using the Bayesian feedback algorithm. The correlation analyses were performed to explore the effect of covariates on the pharmacodynamic parameters. The following variables were tested: the baseline frequencies of SBM or CSBM, age, BMI, and the percentage of female patients in a study. If the P-value of the Pearson correlation was <0.05, the variable was significantly associated with these pharmacodynamic parameters and was selected for the subsequent covariate analyses.21,22 Finally, these selected factors were used to correct the pharmacodynamic parameters using multiple linear regression.

Diagnostic plots were assessed to confirm the model performance. The Monte Carlo method was used to generate 1000 simulations, thus generating 95% confidence interval (CI) for the distribution of drug efficacies. A visual predictive check was then conducted by comparing the 95% CIs with the actual observed values in each drug group to assess the predictive capacity of the final model.

Typical Frequency Change of SBM and CSBM

We performed a single-arm meta-analysis to summarize the pharmacodynamic parameters using a random-effects model. This analysis was performed according to the type of drugs, from which the typical value and the 95% CI of the pharmacodynamic parameters of each drug were obtained. On the basis of these parameters, the typical changes from baseline in the frequencies of SBMs and CSBMs and the 95% CI for each drug at different timepoints were estimated 1000 times using the Monte Carlo simulation. The typical changes in the frequencies of SBMs and CSBMs from baseline values reflect the efficacies of the drugs.

Software

Model establishment and Bayesian feedback were performed using NONMEM7.3 (ICON Development Solutions). Model simulation and generation of plots were performed using R3.3.1 (The R Foundation of Statistical Computing). Meta-analyses were performed using the Stata software, version 13.1 (2013; StataCorp LP, College Station, TX).

RESULTS

Characteristics of Included Studies

A total of 20 studies consisting of 9998 participants were included in the present analysis (Fig. 1). Among these participants, 6459 were in the drug treatment groups, while 3539 were in the placebo groups.23–42 Among the 20 studies included, 6 studies focused on prucalopride (N=1930), 4 on linaclotide (N=1431), 4 on lubiprostone (N=395), 2 on plecanatide (N=1786), 1 on velusetrag (N=294), 1 on bisacodyl (N=247), 1 on sodium picosulfate (N=233), and 1 on elobixibat (N=143). No study focusing on polyethylene glycol, lactulose, or naronapride met the inclusion criteria.

FIGURE 1
FIGURE 1:
Flowchart for study identification and selection. MBMA indicates model-based meta-analysis; RCT, randomized controlled trial.

The mean age of patients ranged from 41.4 to 58.5 years, and the average BMI was 21.5 to 29.8 kg/m2. The mean frequencies of SBM and CSBM at baseline were 1.1 to 4.4 per week and 0.1 to 1.1 per week, respectively. The duration of treatment ranged from 2 to 24 weeks. The detailed information on the included studies has been provided in Table 1. Overall, the included studies were appropriate with a low risk of bias. The details are summarized in Supplemental Digital Content Figure 1 (Supplemental Digital Content 1, http://links.lww.com/JCG/A540).

TABLE 1 - Summary of Included Studies
References Treatment Arms No. Participants Baseline (SBM/CSBM) (Mean±SD) Age (Mean±SD) (y) Gender, Female (%) BMI (Mean±SD) (kg/m2) Duration (wk) Constipation Criteria Primary Outcome
Barish et al23 Placebo 118 1.5±0.80 45.4±13.24 89.0 NA 4 Rome II SBM
Lubiprostone 24 μg 119 1.3±0.88 46.2±12.13 87.4
Fukudo et al26 Placebo 62 1.68±0.77 41.5±14.2 85.5 21.5±2.5 4 Rome III SBM
Lubiprostone 24 μg 62 1.65±0.78 42.7±16.4 90.3 21.5±2.5
Johanson et al28 Placebo 122 1.47±1.33 49.10±12.93 90.2 NA 4 CC SBM
Lubiprostone 24 μg 120 1.37±0.87 48.02±12.28 89.2
Johanson and Ueno29 Placebo 33 1.38 46.8±12.4 93.9 NA 3 CC SBM
Lubiprostone 24 μg 29 1.1 47.6±11.5 96.6
Lubiprostone 48 μg 32 1.35 49.3±12.1 87.5
Lubiprostone 72 μg 33 1.31 49.4±13.8 84.8
DeMicco et al25 Placebo 445 1.55±1.59/0.31±0.50 44.6 78.7 27.97±5.08 12 Rome III SBM/CSBM
Plecanatide 3 mg 443 1.79±2.05/0.28±0.55 45.5 77.9 28.59±4.72
Plecanatide 6 mg 449 1.60±1.66/0.25±0.44 45.3 78.6 28.38±4.86
Miner et al36 Placebo 452 2.2±2.0/0.4±0.6 46.4 79.0 28.1±5.3 12 Rome III SBM/CSBM
Plecanatide 3 mg 453 2.0±1.8/0.3±0.5 45.0 81.2 28.1±5.3
Plecanatide 6 mg 441 1.8±1.8/0.3±0.5 45.1 82.1 28.2±5.3
Johnston et al30 Placebo 10 1.66/0.19 47.3±9.98 70 NA 2 Rome II SBM/CSBM
Linaclotide 100 μg 12 1.80/0.1 48.2±10.20 100
Linaclotide 300 μg 10 2.57/0.26 41.6±15.00 80
Linaclotide 1000 μg 10 1.60/0.32 44.0±13.79 100
Lacy et al33 Placebo 117 1.8±1.2/0.2±0.4 46.4 91.8 28.4±6.1 12 Rome II SBM/CSBM
Linaclotide 145 μg 153 1.7±1.2/0.2±0.4 48.3 90.2 29.2±6.0
Linaclotide 290 μg 159 1.6±1.4/0.2±0.4 47.4 92.5 29.8±6.6
Lembo et al34 Placebo 68 2.3±1.5/0.5±0.6 46.1±15.6 88 NA 4 Rome II SBM/CSBM
Linaclotide 75 μg 59 2.0±1.3/0.3±0.6 46.4±14.2 93
Linaclotide 150 μg 56 2.3±1.5/0.4±0.7 46.4±11.7 96
Linaclotide 300 μg 62 2.2±1.5/0.5±0.7 47.9±12.8 94
Linaclotide 600 μg 62 2.5±1.7/0.3±0.6 49.6±13.7 89
Lembo et al35 Placebo 209 2.0±1.6/0.3±0.6 49 87.1 28±5.4 12 Rome II SBM/CSBM
Linaclotide 145 μg 217 2.1±1.6/0.3±0.6 47 88.0 28±6.5
Linaclotide 290 μg 216 2.0±1.6/0.2±0.4 48 87.0 28±5.3
Placebo 215 1.8±1.4/0.3±0.5 47 91.2 29±7.2
Linaclotide 145 μg 213 1.9±1.5/0.3±0.5 49 91.5 28±5.2
Linaclotide 290 μg 202 1.9±1.6/0.3±0.6 48 88.6 28±5.8
Camilleri et al24 Placebo 209 0.4 48.9 87.6 25.12 12 CC CSBM
Prucalopride 2 mg 207 0.5 48.2 85.3 25.45
Prucalopride 4 mg 204 0.5 47.8 90.8 25.2
Ke et al32 Placebo 252 0.3 41.8±12.88 88.5 22.24 12 CC CSBM
Prucalopride 2 mg 249 0.3 41.4±12.92 91.2 22.56
Piessevaux et al38 Placebo 180 0.43±0.652 48.3±16.25 85.0 24.7±4.36 24 CC CSBM
Prucalopride 2 mg 181 0.38±0.648 49.4±15.78 85.6 25.5±4.80
Quigley et al39 Placebo 212 0.4 46.2 89.2 25.97 12 Rome III CSBM
Prucalopride 2 mg 214 0.4 48.6 84.6 26.12
Prucalopride 4 mg 215 0.5 49.1 86.0 25.26
Tack et al40 Placebo 240 0.4 43.7 92.5 24.5 12 Rome III CSBM
Prucalopride 2 mg 238 0.4 42.7 89.5 24.97
Prucalopride 4 mg 238 0.5 45.4 90.3 24.68
Yiannakou et al41 Placebo 186 0.51 58.5±16.28 0 NA 12 Rome III CSBM
Prucalopride 2 mg 184 0.39 58.4±17.57 0
Goldberg et al27 Placebo 107 1.3±0.86 44.4±11.7 92 27.6±5.4 4 Rome III SBM
Velusetrag 15 mg 101 1.2±0.82 44.3±10.7 93 28.6±6.4
Velusetrag 30 mg 96 1.1±0.80 46.1±10.0 91 27.7±6.5
Velusetrag 50 mg 97 1.2±0.81 45.8±9.3 93 28.0±5.6
Chey et al42 Placebo 47 3.09/0.46 49.9 83.0 26.87±4.330 8 Rome III SBM/CSBM
Elobixibat 5 mg 48 2.82/0.25 48.0 89.6 26.09±4.327
Elobixibat 10 mg 47 2.68/0.50 48.4 95.7 27.09±3.728
Elobixibat 15 mg 48 2.69/0.40 46.2 89.6 27.71±3.841
Kamm et al31 Placebo 121 4.2±2.6/1.0±1.1 54.7±15.1 79.3 26.5±4.7 4 Rome III SBM/CSBM
Bisacodyl 10 mg 247 4.4±3.9/1.1±1.2 55.8±15.9 72.5 26.3±4.5
Mueller-Lissner et al37 Placebo 134 3.2/1.1±0.12 51.9±16.5 85.8 26.1±5.1 4 Rome III SBM/CSBM
Sodium picosulfate 10 mg 233 2.9/0.9±0.09 50.2±17.2 73.0 26.4±4.6
BMI indicates body mass index; CC, chronic constipation; CSBM, complete spontaneous bowel movement; NA, not available; SBM, spontaneous bowel movement.

Model Establishment and Assessment

Thirty-one arms (comprising a total of 4529 participants) reported SBMs, while 31 arms (comprising a total of 5770 subjects) reported CSBMs. The final model described the effect of the drug over time on the frequencies of SBMs and CSBMs. The predicted data was close to the observed data and showed no obvious bias (Supplemental Digital Content Figs. 2, 3, Supplemental Digital Content 1, http://links.lww.com/JCG/A540). The observed changes in the frequencies of SBMs and CSBMs for each drug group were almost within the 95% CI of the predicted data (Figs. 2, 3). This indicates that the model has good predictability.

FIGURE 2
FIGURE 2:
Visual predictive check of the final model of SBM. The points represent observed frequency change of SBM, and symbol size is proportional to sample size. Points linked by a line are from the same arm. Purple lines are the model-predicted fifth, 50th, and 95th percentiles of frequency change of SBM. SBM indicates spontaneous bowel movement.
FIGURE 3
FIGURE 3:
Visual predictive check of the final model of CSBM. The points represent observed frequency change of CSBM, and symbol size is proportional to sample size. Points linked by a line are from the same arm. Purple lines are the model-predicted fifth, 50th, and 95th percentiles of frequency change of CSBM. CSBM indicates complete spontaneous bowel movement.

The parameters Emax,i and τi for each drug group, which were estimated by Bayesian feedback, are provided in Supplemental Digital Content Tables 1 and 3 (Supplemental Digital Content 1, http://links.lww.com/JCG/A540). The correlation analyses showed that the baseline SBM and CSBM frequencies, percentage of female participants, age, and BMI were not significantly associated with the parameters Emax,i or τi. Therefore, there was no need to correct the pharmacodynamic parameters.

Typical Frequency Change of SBM and CSBM

The parameters Emax,i and τi of each drug group were summarized and analyzed according to the type of drug (Tables 2, 3). The results showed that bisacodyl showed the highest Emax value for SBM and CSBM frequencies with values of 6.8 (95% CI: 6.1-7.6) per week and 4.7 (95% CI: 4.3-5.1) per week, respectively. Whereas, plecanatide presented the lowest Emax value for SBM and CSBM frequencies with estimates of 1.6 (95% CI: 1.3-1.9) and 1.0 (95% CI: 0.7-1.2) per week, respectively. These values are approximately one quarter of the values for bisacodyl. In terms of the pharmacodynamic parameter τ, bisacodyl showed the highest loss of efficacy over time. However, plecanatide and elobixibat showed very little loss of efficacy over time, and the 95% CI for their τ values for SBM and CSBM contained 0.

TABLE 2 - Parameter Estimation and Predicted Frequency Change of SBM
Parameter Estimation Predicted Frequency Change Relative to Placebo*
Drugs Arms Sample Size E max (95% CI) τ (95% CI) Week 1 (95% CI) Week 4 (95% CI) Week 12 (95% CI)
Linaclotide 13 1431 2.4 (2.0-2.8) −0.032 (−0.052 to −0.013) 2.3 (1.9-2.7) 2.1 (1.7-2.5) 1.6 (1.2-2.1)
Plecanatide 4 1786 1.6 (1.3-1.9) −0.005 (−0.025 to 0.015) 1.6 (1.3-1.9) 1.6 (1.3-1.9) 1.5 (1.1-2.0)
Lubiprostone 6 395 2.2 (1.9-2.5) −0.047 (−0.083 to −0.011) 2.1 (1.8-2.4) 1.9 (1.5-2.2)
Velusetrag 3 294 2.5 (2.1-2.9) −0.076 (−0.124 to −0.028) 2.3 (1.9-2.7) 1.8 (1.4-2.3)
Elobixibat 3 143 2.4 (1.2-3.7) −0.035 (−0.074 to 0.005) 2.3 (1.1-3.6) 2.1 (1.0-3.3)
Bisacodyl 1 247 6.8 (6.1-7.5) −0.165 (−0.209 to −0.120) 5.8 (5.1-6.5) 3.5 (2.9-4.3)
Sodium picosulfate 1 233 3.8 (3.2-4.4) −0.057 (−0.114 to 0.000) 3.6 (3.0-4.2) 3.0 (2.3-4.0)
Placebo 15 2314 1.7 (1.7-1.8) −0.017 (—)† 1.7 (1.6-1.7) 1.6 (1.5-1.7) 1.4 (1.4-1.5)
*The predicted frequency change of each drug group is the effect after subtracting the corresponding placebo response. The frequency change of each treatment group is predicted on the basis of actual treatment duration.
†The interarm variability of parameter τ of placebo was fixed to 0 to stabilize the model; hence, 95% CI of parameter τ of placebo group cannot be estimated.
CI indicates confidence interval; Emax, theoretical maximal frequency of change from baseline of spontaneous bowel movement; SBM, spontaneous bowel movement; τ, rate of loss of efficacy.

TABLE 3 - Parameter Estimation and Predicted Frequency Change of CSBM
Parameter Estimation Predicted Frequency Change Relative to Placebo*
Drugs Arms Sample Size E max (95% CI) τ (95% CI) Week 1 (95% CI) Week 4 (95% CI) Week 12 (95% CI)
Linaclotide 13 1431 1.5 (1.2-1.7) −0.019 (−0.031 to −0.007) 1.4 (1.2-1.7) 1.4 (1.1-1.6) 1.2 (0.9-1.4)
Plecanatide 4 1786 1.0 (0.7-1.2) 0.001 (−0.007 to 0.009) 1.0 (0.7-1.2) 1.0 (0.7-1.2) 1.0 (0.7-1.2)
Prucalopride 9 1930 1.5 (1.1-1.8) −0.036 (−0.056 to −0.016) 1.4 (1.0-1.8) 1.3 (0.9-1.6) 0.9 (0.6-1.2)
Elobixibat 3 143 1.8 (0.9-2.8) 0.018 (−0.003 to 0.040) 1.9 (0.9-2.8) 2.0 (0.9-3.0)
Bisacodyl 1 247 4.7 (4.3-5.1) −0.159 (−0.193 to −0.126) 4.0 (3.7-4.4) 2.5 (2.1-2.9)
Sodium picosulfate 1 233 2.1 (1.8-2.4) −0.028 (−0.078 to 0.021) 2.0 (1.7-2.3) 1.9 (1.5-2.4)
Placebo 16 3151 0.7 (0.6-0.8) 0.034 (0.030-0.038) 0.7 (0.6-0.8) 0.8 (0.7-0.9) 1.0 (0.9-1.2)
*The predicted frequency change of each drug group is the effect after subtracting the corresponding placebo response. The frequency change of each treatment group is predicted on the basis of actual treatment duration.
CI indicates confidence interval; CSBM, complete spontaneous bowel movement; Emax, theoretical maximal frequency of change from baseline of complete spontaneous bowel movement; τ, rate of loss of efficacy.

On the basis of pharmacodynamic parameters, we predicted the range of distribution for frequency changes in SBMs and CSBMs for each drug at different timepoints (Tables 2, 3). The results showed that the frequency changes in SBMs and CSBMs associated with bisacodyl decreased most significantly over time (Fig. 4). During week 1, bisacodyl increased the frequency of SBMs and CSBMs by 5.8 (95% CI: 5.1-6.5) and 4.0 (95% CI: 3.7-4.4), respectively. However, at week 4, the SBM and CSBM frequencies were reduced to 3.5 (95% CI: 2.9-4.3) and 2.5 (95% CI: 2.1-2.9), respectively. The changes in the frequencies of SBMs and CSBMs associated with plecanatide were the lowest among the drugs at week 1; it increased the frequency of SBMs and CSBMs by only 1.6 (95% CI: 1.3-1.9) and 1.0 (95% CI: 0.7-1.2) per week, respectively. However, these values remained stable at week 4 with 1.5 (95% CI: 1.1-2.0) and 1.0 (95% CI: 0.7-1.2) for SBMs and CSBMs, respectively, due to the low rate of efficacy loss.

FIGURE 4
FIGURE 4:
The predicted typical time course of SBM (A) and CSBM (B) of each drug. The predicted frequency change of SBM and CSBM of each drug is based on the reported treatment duration. CSBM indicates complete spontaneous bowel movement; SBM, spontaneous bowel movement.

Placebo Response

The effect of the placebo on SBMs and CSBMs was also analyzed using the same methods (Supplemental Digital Content Tables 2, 4, Supplemental Digital Content 1, http://links.lww.com/JCG/A540). We found that the baseline frequency of SBMs had a significant effect on the parameter Emax within the placebo group, and the covariate model was expressed as follows:

In Equation 4, Emax,i,placebo,SBM is the Emax value for the ith placebo group on SBMs, and Baselinei is the baseline frequency for SBMs for the ith placebo group. The typical Emax value for the placebo group is 1.72. The correction coefficient of the baseline to the Emax was −0.476. For each reduction in the number of SBMs from the baseline, the Emax value increased by 0.476. N (0, 0.182) represents the intergroup variation of the Emax value using a normal distribution with a mean of 0 and a variance of 0.182.

Similarly, we also estimated the Emax,i and τi values for each placebo group using Bayesian feedback. On the basis of the estimated parameters, the typical effects of the placebo on SBMs and CSBMs were simulated. The results showed that the effect of the placebo on SBM and CSBM did not decrease significantly over time. The frequencies of SBMs and CSBMs increased by 1.40 (95% CI: 1.35-1.45) and 1.0 (95% CI: 0.9-1.2) at week 12, respectively. The placebo effect was comparable to that of the “true effect” of plecanatide; that is, the effect observed in the plecanatide group was approximately twice the magnitude as the effect observed in the placebo group.

DISCUSSION

Chronic constipation is a common gastrointestinal dysfunction disease, and severe cases can cause acute cardiovascular and cerebrovascular events.43,44 However, people are generally unaware as regards the diagnosis and treatment of chronic constipation, thus resulting in delays over the course of chronic constipation.45,46 For patients who failed to achieve symptomatic relief despite increasing daily exercise and consuming high-fiber diets, the use of pharmacotherapy is the next possible treatment option.8,9 However, the descriptions of drug efficacies in current medical guidelines for chronic constipation are mainly qualitative. These guidelines often lack the necessary quantitative information and do not directly reflect the efficacies or characteristics of different drugs. The data show that the frequency of SBMs and CSBMs of some drugs will gradually decrease after a period of treatment. However, the degrees of loss in the frequencies of SBMs and CSBMs has not been compared among different drugs. Therefore, the 2015 guidelines for evaluating medicinal products for the treatment of chronic constipation published by the European Medicines Agency (EMA) suggested that the long-term sustained efficacies of the drugs should be evaluated in clinical trials.13

In this study, we quantitatively evaluated the temporal effectiveness of the 8 drugs used for the treatment of chronic constipation by establishing pharmacodynamic models. We found that bisacodyl had the highest effect on increasing the frequencies of SBMs and CSBMs; this effect is ∼4 times greater than that of plecanatide. This conclusion is consistent with a study published in Gut in 2017.14 However, the duration of treatment for clinical trials on bisacodyl that were included in this study was only 4 weeks. Furthermore, the safety and efficacy of long-term bisacodyl treatment had not been evaluated. Although a previous study found that, except for occasionally causing abdominal pain, bisacodyl was well-tolerated for short-term use, it is necessary to know whether tolerance develops with more-prolonged therapy. This could help clinicians determine whether intermittent treatment might produce a more-sustained and better-tolerated effect for some patients. In addition, this study found that bisacodyl had the highest rate of efficacy loss, with the frequencies of SBMs and CSBMs decreasing by about 40% between week 1 and week 4. In other words, the long-term use of bisacodyl cannot guarantee its safety, and it also loses its efficacy over time. Therefore, it is recommended that patients who need to increase their increased frequency of bowel movements to relieve constipation as soon as possible should consider bisacodyl for short-term use.

Plecanatide and linaclotide have the same mechanism of action and belong to the guanylate cyclase-C agonist group.47,48 A recent meta-analysis showed no differences in efficacy or the frequency of adverse events between plecanatide and linaclotide.49 However, this study also showed that the increases in the frequencies of SBMs and CSBMs during treatment with plecanatide were significantly lower than that of linaclotide at week 1. However, there was no significant difference at week 12 due to the lower rate of efficacy loss observed with plecanatide. Therefore, compared with the previous meta-analysis, this study distinguishes the differences in the efficacies over time for different drugs. In terms of safety, linaclotide and plecanatide carry the potential risk of dehydration and diarrhea. Thus, the US Food and Drug Administration (FDA) points out that these 2 drugs should be avoided in nonadult patients.50 In addition, previous studies have shown that the dropout rate from plecanatide trials due to severe diarrhea is much lower than that of linaclotide. This finding indicates that plecanatide is relatively well-tolerated in patients.50,51

The results of this study revealed that there were no significant differences in the increased frequencies of SBMs and CSBMs when using prucalopride, velusetrag, lubiprostone, or elobixibat. Among these drugs, prucalopride and lubiprostone were approved by the FDA for the treatment of chronic idiopathic constipation.52,53 Prucalopride is available in Europe, and the World Gastrointestinal Organization guidelines for constipation listed this drug as a level A recommendation.4,54 Velusetrag and elobixibat are still under development.27,55 Previous studies have shown that these 4 drugs are safe to use, and their main adverse effects include headache, nausea, diarrhea, and abdominal pain.23,24,27,29,55

The frequencies of SBMs and CSBMs per week are some of the most commonly used indicators in clinical trials for treating chronic constipation. An SBM is defined as stool not induced by rescue medication, whereas, a CSBM is defined as an SBM that is associated with a sensation of complete evacuation. A guideline published by the EMA indicates that a CSBM is more sensitive than an SBM in clinical trials on chronic constipation.13 This study shows that the changes in the frequencies of SBMs and CSBMs between baseline and week 4 were very low in the elobixibat group, and the difference between these 2 indicators was 0.2 times per week. Whereas, the difference between the SBM and CSBM frequencies in the linaclotide and plecanatide groups at week 4 was ∼0.7 times per week. The difference between SBM and CSBM frequencies in bisacodyl and sodium picosulfate groups was the most obvious, with differences of 1.0 and 1.1 times per week, respectively. The above results suggest that the sensation of defecation differed among treatments with different drugs. The bowel movements caused by elobixibat would be classified as CSBMs, whereas some of the bowel movements caused by bisacodyl and sodium picosulfate would be classified as non-CSBMs. Therefore, elobixibat is superior to bisacodyl and sodium picosulfate by producing the sensation of complete evacuation. Elobixibat is an inhibitor of the ileal bile acid transporter, which increases the bile acid concentration in the gut and leads to increases in fluid secretion, motility, and the resultant acceleration of transit.56 Bisacodyl and sodium picosulfate were used as stimulant laxatives, meaning that they work directly on the colon to produce a bowel movement.57,58 It has been hypothesized that these differences in the mechanism of action may produce a better sense of use of elobixibat.

We found that the change in frequency of bowel movements after intervention was not associated with the baseline frequencies of SBMs and CSBMs, BMI, age, and percentage of female patients in a study, which might be related to the narrow ranges of the baseline of patients included in the study. In addition, we also explored the dose-effect relationship on the change in the frequency of bowel movements. The results showed that there was little change in the frequency of bowel movements within the dose range of 24 to 72 μg for lubiprostone, 3 to 6 mg for plecanatide, 75 to 1000 µg for linaclotide, 15 to 50 mg for velusetrag, and 2 to 4 mg for prucalopride. Elobixibat exhibited a significant dose-response relationship in the range of 5 to 15 mg, but this conclusion was based on a trial with a small sample size. Therefore, the dose-response relationship of elobixibat requires further investigation. The above results suggest that a higher dose of most drugs does not significantly increase the frequencies of SBMs and CSBMs compared with lower doses. To ensure medication safety, low doses of drugs are usually recommended in clinical practice. For example, the FDA-approved lubiprostone at a dose of 24 µg twice a day, plecanatide at a dose of 3 mg once daily, and linaclotide at a dose of 145 µg once daily. Similarly, the EMA approved prucalopride at a dose of 2 mg once per day for the treatment of chronic idiopathic constipation.

In this study, we also quantified the features of the response to the placebo in the included studies. The results showed that the placebo effect at week 12 was comparable with the response to placebo at week 1, thus indicating that the placebo effect did not decrease with time. The baseline SBM frequency was significantly associated with the Emax value of SBM. Because of the narrow distribution of the CSBM baseline values (range: 0.2 to 1.1/wk) in the placebo groups, the baseline frequency of CSBMs was not found to be associated with the placebo effect of CSBM.

There were several limitations to this study. First, only 1 trial each for bisacodyl and sodium picosulfate was included in our analysis. Therefore, the efficacy estimates for these 2 drugs may be biased by sampling error. However, this study was carried out to quantify the true drug effects by deducting the placebo effects, thereby reducing the heterogeneity among studies due to differences in patient exercise and diet control. In general, variations in the size of the true drug effect among studies were not very large. Second, the duration of treatment for some drugs was only 4 weeks. Although the established model in this study can predict the change in the frequencies of SBMs and CSBMs under long-term drug treatment, the reliability of the extrapolated results after 4 weeks requires further verification through clinical trials. Moreover, as this study only focused on the quantitative evaluation of drug efficacy on relieving constipation, adverse drug reactions were not analyzed due to limitations associated with available data from the literature. However, we have outlined the common adverse reactions for each drug in the discussion section. Finally, there is potential for publication bias, as we only included studies published in English.

CONCLUSIONS

This study quantitatively evaluated the features of changes in the frequency of bowel movements associated with 8 drugs for the treatment of chronic constipation. Bisacodyl showed the greatest increase in the frequency of bowel movements. However, it also showed the greatest loss in efficacy over time. Plecanatide showed the smallest increase in the frequency of bowel movements, but its effect on the frequencies of SBMs and CSBMs did not decrease over time. This quantitative study supplements the available guidelines on chronic constipation pharmacotherapy.

REFERENCES

1. Brandt LJ, Prather CM, Quigley EM, et al. Systematic review on the management of chronic constipation in North America. Am J Gastroenterol. 2005;100(suppl 1):S5–S21.
2. Heidelbaugh JJ, Stelwagon M, Miller SA, et al. The spectrum of constipation-predominant irritable bowel syndrome and chronic idiopathic constipation: US survey assessing symptoms, care seeking, and disease burden. Am J Gastroenterol. 2015;110:580–587.
3. Higgins PD, Johanson JF. Epidemiology of constipation in North America: a systematic review. Am J Gastroenterol. 2004;99:750–759.
4. Lindberg G, Hamid SS, Malfertheiner P, et al. World Gastroenterology Organisation Global Guideline: constipation—a global perspective. J Clin Gastroenterol. 2011;45:483–487.
5. Anti M, Pignataro G, Armuzzi A, et al. Water supplementation enhances the effect of high-fiber diet on stool frequency and laxative consumption in adult patients with functional constipation. Hepatogastroenterology. 1998;45:727–732.
6. Bove A, Bellini M, Battaglia E, et al. Consensus statement AIGO/SICCR diagnosis and treatment of chronic constipation and obstructed defecation (part II: treatment). World J Gastroenterol. 2012;18:4994–5013.
7. Ternent CA, Amir LB, Morin NA, et al. Practice parameters for the evaluation and management of constipation. Dis Colon Rectum. 2007;50:2013–2022.
8. American Gastroenterological Association, Bharucha AE, Dorn SD, Lembo A, et al. American Gastroenterological Association medical position statement on constipation. Gastroenterology. 2013;144:211–217.
9. Mearin F, Ciriza C, Mínguez M, et al. Clinical practice guideline: irritable bowel syndrome with constipation and functional constipation in the adult. Rev Esp Enferm Dig. 2016;108:332–363.
10. Johnson DA. Treating chronic constipation: how should we interpret the recommendations? Clin Drug Investig. 2006;26:547–557.
11. Ramkumar D, Rao SS. Efficacy and safety of traditional medical therapies for chronic constipation: systematic review. Am J Gastroenterol. 2005;100:936–971.
12. Paquette IM, Varma M, Ternent C, et al. The American Society of Colon and Rectal Surgeons Clinical Practice Guideline for the evaluation and management of constipation. Dis Colon Rectum. 2016;59:479–492.
13. European Medicines Agency. Guideline on the evaluation of medicinal products for the treatment of chronic constipation (including opioid induced constipation) and for bowel cleansing; 2015.
14. Nelson AD, Camilleri M, Chirapongsathorn S, et al. Comparison of efficacy of pharmacological treatments for chronic idiopathic constipation: a systematic review and network meta-analysis. Gut. 2017;66:1611–1622.
15. Boucher M, Bennetts M. The many flavors of model-based meta-analysis:part I: introduction and landmark data. CPT Pharmacometrics Syst Pharmacol. 2016;5:54–64.
16. Mandema JW, Salinger DH, Baumgartner SW, et al. A dose response meta-analysis for quantifying relative efficacy of biologics in rheumatoid arthritis. Clin Pharmacol Ther. 2011;90:828–835.
17. Pedder H, Dias S, Bennetts M, et al. Modelling time-course relationships with multiple treatments: model-based network meta-analysis for continuous summary outcomes. Res Synth Methods. 2019;10:267–286.
18. Li L, Xu L, Wu J, et al. Comparative efficacy of nonhormonal drugs on menopausal hot flashes. Eur J Clin Pharmacolo. 2016;72:1051–1058.
19. Checchio T, Ahadieh S, Gupta P, et al. Quantitative evaluations of time-course and treatment effects of systemic agents for psoriasis: a model-based meta-analysis. Clin Pharmacol Ther. 2017;102:1006–1016.
20. Dong Z, Xu L, Liu H, et al. Comparative efficacy of five long-term weight loss drugs: quantitative information for medication guidelines. Obes Rev. 2017;18:1377–1385.
21. Gibbs JP, Fredrickson J, Barbee T, et al. Quantitative model of the relationship between dipeptidyl peptidase-4 (DPP-4) inhibition and response: meta-analysis of alogliptin, saxagliptin, sitagliptin, and vildagliptin efficacy results. J Clin Pharmacol. 2012;52:1494–1505.
22. Wählby U, Jonsson EN, Karlsson MO. Assessment of actual significance levels for covariate effects in NONMEM. J Pharmacokinet Phar. 2001;28:231–252.
23. Barish CF, Drossman D, Johanson JF, et al. Efficacy and safety of lubiprostone in patients with chronic constipation. Dig Dis Sci. 2010;55:1090–1097.
24. Camilleri M, Kerstens R, Rykx A, et al. A placebo-controlled trial of prucalopride for severe chronic constipation. N Engl J Med. 2008;358:2344–2354.
25. DeMicco M, Barrow L, Hickey B, et al. Randomized clinical trial: efficacy and safety of plecanatide in the treatment of chronic idiopathic constipation. Therap Adv Gastroenter. 2017;10:837–851.
26. Fukudo S, Hongo M, Kaneko H, et al. Lubiprostone increases spontaneous bowel movement frequency and quality of life in patients with chronic idiopathic constipation. Clin Gastroenterol Hepatol. 2015;13:294.e5–301.e5.
27. Goldberg M, Li YP, Johanson JF, et al. Clinical trial: the efficacy and tolerability of velusetrag, a selective 5-HT4 agonist with high intrinsic activity, in chronic idiopathic constipation—a 4-week, randomized, double-blind, placebo-controlled, dose-response study. Aliment Pharmacol Ther. 2010;32:1102–1112.
28. Johanson JF, Morton D, Geenen J, et al. Multicenter, 4-week, double-blind, randomized, placebo-controlled trial of lubiprostone, a locally-acting type-2 chloride channel activator, in patients with chronic constipation. Am J Gastroenterol. 2008;103:170–177.
29. Johanson JF, Ueno R. Lubiprostone, a locally acting chloride channel activator, in adult patients with chronic constipation: a double-blind, placebo-controlled, dose-ranging study to evaluate efficacy and safety. Aliment Pharmacol Ther. 2007;25:1351–1361.
30. Johnston JM, Kurtz CB, Drossman DA, et al. Pilot study on the effect of linaclotide in patients with chronic constipation. Am J Gastroenterol. 2009;104:125–132.
31. Kamm MA, Mueller-Lissner S, Wald A, et al. Oral bisacodyl is effective and well-tolerated in patients with chronic constipation. Clin Gastroenterol Hepatol. 2011;9:577–583.
32. Ke M, Zou D, Yuan Y, et al. Prucalopride in the treatment of chronic constipation in patients from the Asia-Pacific region: a randomized, double-blind, placebo-controlled study. Neurogastroenterol Motil. 2012;24:999.e541.
33. Lacy BE, Schey R, Shiff SJ, et al. Linaclotide in chronic idiopathic constipation patients with moderate to severe abdominal bloating: a randomized, controlled trial. PLoS One. 2015;10:e0134349.
34. Lembo AJ, Kurtz CB, Macdougall JE, et al. Efficacy of linaclotide for patients with chronic constipation. Gastroenterology. 2011;138:886–895.
35. Lembo AJ, Schneier HA, Shiff SJ, et al. Two randomized trials of linaclotide for chronic constipation. N Engl J Med. 2011;365:527–536.
36. Miner PB Jr, Koltun WD, Wiener GJ, et al. A randomized phase III clinical trial of plecanatide, a uroguanylin analog, in patients with chronic idiopathic constipation. Am J Gastroenterol. 2017;112:613–621.
37. Mueller-Lissner S, Kamm MA, Wald A, et al. Multicenter, 4-week, double-blind, randomized, placebo-controlled trial of sodium picosulfate in patients with chronic constipation. Am J Gastroenterol. 2010;105:897–903.
38. Piessevaux H, Corazziari E, Rey E, et al. A randomized, double-blind, placebo-controlled trial to evaluate the efficacy, safety, and tolerability of long-term treatment with prucalopride. Neurogastroenterol Motil. 2015;27:805–815.
39. Quigley EM, Vandeplassche L, Kerstens R, et al. Clinical trial: the efficacy, impact on quality of life, and safety and tolerability of prucalopride in severe chronic constipation-a 12-week, randomized, double-blind, placebo-controlled study. Aliment Pharmacol Ther. 2009;29:315–328.
40. Tack J, Van Outryve M, Beyens G, et al. Prucalopride (Resolor) in the treatment of severe chronic constipation in patients dissatisfied with laxatives. Gut. 2009;58:357–365.
41. Yiannakou Y, Piessevaux H, Bouchoucha M, et al. A randomized, double-blind, placebo-controlled, phase 3 trial to evaluate the efficacy, safety, and tolerability of prucalopride in men with chronic constipation. Am J Gastroenterol. 2015;110:741–748.
42. Chey WD, Camilleri M, Chang L, et al. A randomized placebo-controlled phase IIb trial of A3309, a bile acid transporter inhibitor, for chronic idiopathic constipation. Am J Gastroenterol. 2011;106:1803–1812.
43. Dennison C, Prasad M, Lloyd A, et al. The health-related quality of life and economic burden of constipation. Pharmacoeconomics. 2005;23:461–476.
44. Irvine EJ, Ferrazzi S, Pare P, et al. Health-related quality of life in functional GI disorders: focus on constipation and resource utilization. Am J Gastroenterol. 2002;97:1986–1993.
45. Stewart WF, Liberman JN, Sandler RS, et al. Epidemiology of constipation (EPOC) study in the United States: relation of clinical subtypes to sociodemographic features. Am J Gastroenterol. 1999;94:3530–3540.
46. Wald A, Scarpignato C, Mueller-Lissner S, et al. A multinational survey of prevalence and patterns of laxative use among adults with self-defined constipation. Aliment Pharmacol Ther. 2008;28:917–930.
47. Shailubhai K, Comiskey S, Foss JA, et al. Plecanatide, an oral guanylate cyclase C agonist acting locally in the gastrointestinal tract, is safe and well-tolerated in single doses. Dig Dis Sci. 2013;58:2580–2586.
48. Busby RW, Bryant AP, Bartolini WP, et al. Linaclotide, through activation of guanylate cyclase C, acts locally in the gastrointestinal tract to elicit enhanced intestinal secretion and transit. Eur J Pharmacol. 2010;649:328–335.
49. Shah ED, Kim HM, Schoenfeld P. Efficacy and tolerability of guanylate cyclase-C agonists for irritable bowel syndrome with constipation and chronic idiopathic constipation: a systematic review and meta-analysis. Am J Gastroenterol. 2018;113:329–338.
50. Islam BN, Sharman SK, Browning DD. Clinical utility of plecanatide in the treatment of chronic idiopathic constipation. Int J Gen Med. 2018;11:323–330.
51. Martinez de Andino N. Current treatment paradigm and landscape for the management of chronic idiopathic constipation in adults: focus on plecanatide. J Am Assoc Nurse Pract. 2018;30:412–420.
52. Vijayvargiya P, Camilleri M. Use of prucalopride in adults with chronic idiopathic constipation. Expert Rev Clin Pharmacol. 2019;23:1–11.
53. McKeage K, Plosker GL, Siddiqui MA. Lubiprostone. Drugs. 2006;66:873–879.
54. Camilleri M, Deiteren A. Prucalopride for constipation. Expert Opin Pharmacother. 2010;11:451–461.
55. Miner PB. Elobixibat, the first-in-class ileal bile acid transporter inhibitor, for the treatment of chronic idiopathic constipation. Expert Opin Pharmacother. 2018;19:1381–1388.
56. Acosta A, Camilleri M. Elobixibat and its potential role in chronic idiopathic constipation. Ther Adv Gastroenterol. 2014;7:167–175.
57. Manabe N, Cremonini F, Camilleri M, et al. Effects of bisacodyl on ascending colon emptying and overall colonic transit in healthy volunteers. Aliment Pharmacol Ther. 2009;30:930–936.
58. Kienzle-Horn S, Vix JM, Schuijt C, et al. Comparison of bisacodyl and sodium picosulfate in the treatment of chronic constipation. Curr Med Res Opin. 2007;23:691–699.
Keywords:

chronic constipation; pharmacotherapy; spontaneous bowel movement; complete spontaneous bowel movement; loss of efficacy; model-based meta-analysis

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.