An elephant in a dark room: Inflammatory bowel disease from the molecule to the three-minute consultation

Abstract

In Book III of the Masnavi, Rumi tells a story of an elephant brought from India and kept in a dark house. People came to see it, and because they could see nothing, each felt the animal with the palm of a hand. The one who reached the trunk said the creature was a water-pipe; the one who touched an ear said it was a fan; the one who took hold of a leg said a pillar; the one who ran a hand along the back said a throne. The eye of sense-perception, Rumi writes, is only like the palm of the hand, and the palm has no means of reaching the whole. Had each man carried a candle, he adds, the differences would have gone out of their words [Figure 1].

Figure 1.

The story of an elephant brought from India and kept in a dark house and different perceptions and interpretations of various individuals based on their sensory inputs; although each of them is correct, none has a complete picture

The five articles on inflammatory bowel disease (IBD) in this issue of the Saudi Journal of Gastroenterology are five hands on the same animal.[1,2,3,4,5] One counts it, the other takes it apart at the level of a cytokine, while another describes the symptom patients rank as the most disabling. Then, a fourth asks what the disease does to a person’s standing among other people. Finally, a fifth records what actually happens in the clinic room. Each is careful within its own frame. None of them, alone, is about IBD [Figure 2, Table 1].

Figure 2.

Five hands on the same animal. Each of the five inflammatory bowel disease articles in this issue reports a different part of the same condition — its burden, its molecular target, its dominant symptom, its social cost, and the consultation in which all of these are supposed to meet. None of them, alone, describes the whole

Table 1.

The five inflammatory bowel disease articles in this issue at a glance

Article	Design and Sample	Principal Finding	Chief Constraint
Jung et al. GBD Asia	Modeled analysis of GBD 2023 estimates; 34 countries, 5 Asian regions, 1990–2023	Age-standardized prevalence rose and age-standardized DALY rates fell in every region; Central Asia highest ASDR (17.08 per 100,000), South Asia highest ASPR (48.83 per 100,000)	Estimates are model outputs, not measurements; West Asia is outside the regional frame
Choi et al. TL1A inhibitors	Systematic review of all phase 2 data; 6 studies from 5 trials, 804 patients	Ulcerative colitis remission RR 2.56 (1.60–4.09), NNT 4.5; Crohn’s endoscopic response RR 3.67 (1.64–8.20) with duvakitug 900 mg; GRADE moderate	A 1% placebo remission rate in one cohort; I2=0% across only three trials; two single-arm studies at serious risk of bias
Alruwaili et al. Fatigue	Narrative review with structured literature synthesis, 2010–2024	About half of unselected adults report clinically significant fatigue, persisting in remission and correlating weakly with CRP, calprotectin, and endoscopic activity	Almost all therapeutic fatigue data are secondary or post hoc; no Arabic-validated instrument in routine use
Alzahrani Stigma and support	Cross-sectional online survey; 103 Saudi adults with self-reported IBD	Concealment strategies associated with lower perceived support and poorer quality of life; family the highest-rated support source; 86% reported recent negative emotions	Convenience and snowball sampling, self-reported diagnosis, no disease activity data; association, not mediation
Al Sulais et al. Shared decision-making	Multicenter cross-sectional study; 105 audio-recorded encounters, 3 centers	Median OPTION-5 score 5 of 100 (each component median 0); median consultation 3 minutes; 83% of patients nonetheless completely agreed the decision was shared (ρ=−0.044)	Three clinicians in one province; inter-rater reliability not reported; three-quarters were routine follow-ups

ASDR=age-standardized disability-adjusted life-year rate, ASPR=age-standardized prevalence rate, CRP=C-reactive protein, DALY=disability-adjusted life-year, GBD=Global Burden of Disease, IBD=inflammatory bowel disease, NNT=number needed to treat, RR=risk ratio

THE HAND ON THE FLANK: HOW MANY, AND WHERE

Jung and colleagues apply the Global Burden of Disease (GBD) Study 2023 to 34 countries in five Asian regions.[1] Their central finding is a pair of scissors: between 1990 and 2023 the age-standardized prevalence rate (ASPR) rose in every region while the age-standardized disability-adjusted life-year rate (ASDR) fell in every region [Figure 3]. East Asia shows both blades most sharply; ASDR average annual percentage change (AAPC) −2.32; ASPR AAPC + 1.50. In 2023, Central Asia carried the highest regional ASDR at 17.08 per 100,000, South Asia the highest ASPR at 48.83 per 100,000, and the Republic of Korea the highest national ASPR at 57.70 per 100,000.

Figure 3.

The scissors: prevalence up, disability down. Average annual percentage change in age-standardized prevalence rate and age-standardized disability-adjusted life-year rate for overall inflammatory bowel disease across the five Asian regions of the Global Burden of Disease Study 2023, 1990–2023, as reported by Jung and colleagues in this issue. Bars show point estimates; horizontal lines show 95% confidence intervals. West Asia is not included in this regional framework

There are two ways to read those scissors, and both are probably true. The optimistic reading is the one we have been anticipating since Ng and colleagues assembled the population-based data[6] and Kaplan and Windsor described compounding prevalence as the later stages in the global evolution of IBD[7]: more people are living with the disease, and each of them carries less measured disability, because diagnosis comes earlier and treatment is better. The sober reading is that a disability-adjusted life-year is a construct rather than an observation. The authors are commendably direct about this: ASDR declined even in settings such as Kyrgyzstan and Tajikistan where access to biologic therapy is minimal, which they interpret as evidence that disability-weight specification and model-based estimation in data-sparse settings are doing part of the work. A falling ASDR in a country with almost no infliximab is not a clinical result; it is a modeling result, and it deserves to be read as one. Successive GBD cycles have put the global prevalent case count at 6.8 million in 2017 and around 4.9 million in 2019; these estimates differ chiefly because the model was revised, not because the disease receded.[8,9]

Then there is the hand that never reached the animal. The GBD architecture used here defines Asia as Central, East, South, and Southeast Asia together with high-income Asia Pacific. Mauritius and Seychelles are included, on the defensible grounds of Indian Ocean history. West Asia, the Arabian Peninsula, is not in the frame at all. That is a property of GBD regionalization rather than a fault of the authors, but the consequence for us is concrete: a reader in Riyadh or Jeddah cannot find their own population on a map of their own disease.

Thirteen years ago, in this journal, one of us wrote with Subrata Ghosh that IBD had become a global disease and that Saudi Arabia had the economic resources to invest in IBD research, including genetics, environment, and microbiome.[10] Since then, we have produced national consensus guidance for adults and for children with IBD;[11,12] what we have not produced is a denominator. We can now say with reasonable confidence what to prescribe, but we still cannot say to how many.

THE HAND ON THE TUSK: THE MOLECULE

Choi et al.[2] systematically review phase 2 data for the three anti-TL1A monoclonal antibodies: Tulisokibart, Afimkibart, and Duvakitug in 804 patients. TL1A, encoded by TNFSF15, is an attractive target for a reason beyond adding another anti-inflammatory mechanism: TNFSF15 variants have been linked to IBD susceptibility by genome-wide association, and the TL1A–DR3 axis activates intestinal fibroblasts directly, independently of inflammation. If that holds in humans, it would address the fibrostenotic aspect of the disease that drives strictures and resection.

The pooled numbers are good; in ulcerative colitis, the risk ratio versus placebo was 2.56 (95% CI 1.60 to 4.09) for clinical remission and 2.13 (95% CI 1.44 to 3.16) for endoscopic improvement, with risk differences of +22.2 and +22.6 percentage points and numbers needed to treat of 4.5 and 4.4. In Crohn’s disease, Duvakitug 900 mg achieved an endoscopic response risk ratio of 3.67 (95% CI 1.64 to 8.20), while GRADE certainty was moderate for the ulcerative colitis outcomes [Figure 4]. The denominators show something interesting: clinical remission in the placebo arm of ARTEMIS-UC was 1%, against the 6–15% usual in contemporary ulcerative colitis trials.[13,14,15,16] A number needed to treat is an inverted risk difference, and a risk difference is only as trustworthy as the arm it is subtracted from. Second, an I² of 0.0% across three trials of three structurally distinct antibodies is being asked to carry the phrase “class effect.” With three studies, I² is a statement about statistical power rather than about biology, a caution the authors offer and that we would underline. Third, two of the six studies were single-arm with historical comparators and were rated at a serious risk of bias, and one randomized trial used a nine-arm allocation with small per-arm numbers and a primary endpoint different from the score regulators prefer.[15]

Figure 4.

Promising and provisional: anti-TL1A antibodies in phase 2. Pooled phase 2 risk ratios versus placebo reported by Choi and colleagues in this issue, with risk differences and numbers needed to treat. Diamonds show point estimates and horizontal lines 95% confidence intervals on a logarithmic scale. The two caution panels identify the features of the evidence base that most constrain interpretation. NNT = number needed to treat, pp = percentage points, RR = risk ratio, UC = ulcerative colitis

The most interesting claim is the one these data cannot yet address. An anti-fibrotic effect is a question about strictures, obstruction, and resection over years; it is currently being examined with endoscopic surrogate endpoints at week 12. That mismatch is not a criticism of the trials but rather a description of where the field stands. The question for our health system will not be whether these agents work, it will be for whom, in what sequence, and at what price. We already position a growing list of advanced therapies with almost no head-to-head data, which is why our own guidance, like others, endorses several acceptable first-line options.[11] Another effective molecule does not resolve that problem; it enlarges it.

THE HAND ON THE EAR: FATIGUE

Alruwaili et al.[3] review fatigue in IBD. We should declare an interest here rather than bury it: both of us are authors of that review, and readers should discount our enthusiasm accordingly. Having previously argued in these pages that disclosure is a weak instrument,[17] we are at least obliged to use it.

The substance is uncomfortable, approximately half of unselected adults with IBD report clinically significant fatigue, and a substantial proportion continue to report it in clinical or biochemical remission. Fatigue tracks more closely with mood, sleep quality, and subjective disease burden than with C-reactive protein, fecal calprotectin, or endoscopic activity. The symptom patients most often name as the worst part of the disease is largely invisible to the instruments on which we have built treat-to-target care.

STRIDE-II recognized patient-reported outcomes, fatigue among them, as long-term treatment targets five years ago.[16] Nonetheless, fatigue is almost never a primary endpoint in any pivotal trial. Essentially everything we tell patients about fatigue and advanced therapy comes from prespecified secondary analyses or post hoc exploration, and the most useful thing this review does is to say so explicitly, evidence level by evidence level, instead of blending the three. It is also honest about the ruler: a clinically meaningful change on the FACIT-Fatigue scale is context-specific, with roughly 6 to 10 points having the strongest empirical support across IBD studies. We have been declaring victories against a moving threshold.

For our region there is an additional and unglamorous task. Instruments validated in English do not travel freely into Arabic and translation is not validation. If fatigue is to become a target rather than a topic, someone has to do the psychometric work, and then someone must build the resulting question into a clinical workflow.

THE HAND ON THE FOOT: STIGMA, AND THE FAMILY

Alzahrani surveyed 103 Saudi adults with a self-reported gastroenterologist diagnosis of Crohn’s disease or ulcerative colitis, recruited through online patient groups and messaging networks.[4] Education and advocacy were the most frequently used stigma management strategies, and withdrawal was the least. Family was the highest-rated source of perceived social support. Concealment-based strategies were associated with lower perceived support and, through it, with poorer quality of life across all WHOQOL-BREF domains; the indirect association with overall quality of life was − 0.11. About 86% of participants reported negative emotions sometimes, often, or always in the preceding four weeks.

This is a cross-sectional convenience and snowball sample of 103 people, predominantly young and predominantly women, with a self-reported diagnosis and no measure of disease activity. Bootstrapped indirect paths estimated in such a sample describe association, not mediation in any causal sense, and the qualifiers around the word “mediation” tend to fall away as a finding travels. What the study does establish is a direction consistent with a substantial international literature on internalized stigma in IBD,[1,18,19] now with Saudi participants rather than by extrapolation. That is worth having.

The family finding deserves separate attention. In a setting where a companion attends roughly a third of gastroenterology visits, the family is not background context, it is an underused therapeutic asset that we neither assess nor systematically enlist.

THE HAND ON THE TRUNK: THE CONSULTATION

Al Sulais et al. audio-recorded gastroenterology encounters at three centers in the Eastern Province and scored them with the observer-based OPTION-5 instrument,[20] while patients rated the same encounters with the 9-item Shared Decision-Making Questionnaire.[21] The results are the most arresting in this issue [Figure 5]. Across 105 evaluable encounters, the median OPTION-5 score was 5 out of 100 (interquartile range 0 to 10), and the median score for each of the five individual components was 0; a published synthesis of 29 studies reports a mean OPTION-5 of 23.[22] Longer consultations and the decision to start a new medication independently predicted higher scores.

Figure 5.

The gap between the consultation and the memory of it. Observer-rated (OPTION-5) and patient-reported (SDM-Q-9) shared decision-making in the same 105 inflammatory bowel disease encounters, as reported by Al Sulais and colleagues in this issue. The two measures were uncorrelated. The reference value of 23 is the mean OPTION-5 score across 29 previously published studies

The median consultation lasted three minutes and yet 83% of patients completely agreed that the decision had been made in complete agreement with their physician, and 70% that treatment options had been discussed. The correlation between what the recording contained and what the patient remembered was ρ = −0.044 (P = 0.687). We do not read this as patients being wrong but as calibration. People rate an encounter against what they expected of it, and a courteous, efficient three-minute visit with a physician they trust meets that expectation completely. Satisfaction is not participation, and a system that has never offered the alternative cannot expect to be marked down for failing to offer it.

There are honest limits here too: three clinicians, one province, no reported inter-rater reliability, and three-quarters of visits were routine follow-ups at which no decision was on the table, a situation in which OPTION-5 has a floor. The finding survives all of them, because the effect is not marginal. A median of 5 out of 100 is not measurement noise. The remedy is not exhortation; ten-minute appointment slots that become three-minute encounters are a design output, not a character trait; every system is designed to get the results it gets.

THE CANDLE

Set the five hands side by side and what is absent is more striking than what is present. We have global modelled estimates and no local denominator. We have a symptom that dominates patients’ lives and no validated instrument in their language. We have evidence that concealment costs quality of life and no service that asks about it. We have a lengthening catalog of expensive molecules and three minutes in which to choose among them.

Four things would carry a candle into the room, and none of them requires a discovery. First, a national IBD registry with a genuine population denominator, capable of feeding the global estimates so that West Asia stops being inferred from elsewhere. Second, Arabic-validated patient-reported outcomes for fatigue and quality of life, captured before the visit rather than during it, which returns minutes to the consultation. Third, consultation design: protected time when a therapeutic decision is due, decision aids, and shared decision-making measured rather than assumed. Fourth, comparative effectiveness and health technology assessment ahead of formulary decisions, so that positioning is determined by evidence and value rather than by order of arrival. We have made a version of this argument before about colorectal cancer screening: offering the intervention as a small part of the story, and doing the right thing is not the same as doing it right.[23]

Rumi ends the passage by observing that had each of the men in that dark house carried a candle, the difference would have gone out of their words. The five hands in this issue are all on the same animal. The candle is not another molecule. It is a denominator, a validated question asked in the patient’s own language, and enough minutes to hear the answer.