This prospective cohort examined whether habitual qat chewing influences short-term response to routine conservative care for temporomandibular disorders (TMD) in Sana’a, Yemen. The investigators enrolled consecutive, treatment-seeking patients diagnosed by RDC/TMD Axis I and reassessed them after three months of usual care. The question asked was straightforward: do habitual chewers fare worse than non-chewers when managed conservatively?
Consecutive patients presenting with signs or symptoms of TMD were screened between October 2018 and October 2020. A total of 156 were screened; 148 met eligibility and were enrolled following RDC/TMD Axis I examination. Habitual chewing was defined as self-reported qat use in three or more sessions per week. Fifty-four participants (36.5%) met that threshold and 94 did not. Every participant attended the three-month follow-up visit; there were no losses to follow-up and no missing outcome data.
Diagnosis and all examinations used the RDC/TMD Axis I protocol rather than the later DC/TMD. A single experienced oral medicine specialist performed every baseline and follow-up assessment after calibration, with inter-examiner agreement reported prior to recruitment (Cohen’s kappa >0.85 for key dichotomous items). The same examiner recorded exposure and outcome, so assessments were not blinded. The study team acknowledged that, within this clinical setting, concealment of qat use was unrealistic; they also noted that any examiner expectation that chewers do worse would bias results toward an exaggerated disadvantage for chewers rather than masking one.
All patients received reassurance, education and a self-care programme including soft diet, avoidance of wide opening and parafunctional loading, and local heat or cold. Treating clinicians added pharmacotherapy or a stabilization splint according to clinical need. Drug classes used were NSAIDs (naproxen or diclofenac sodium), diazepam, and amitriptyline; patients received either single agents or combinations. A stabilization splint was supplied when parafunctional signs or persistent joint loading were present.
Key treatment details were not transcribed into the research dataset. Prescribed doses, dosing frequency, intended duration, splint design features and structured safety data were recorded in clinical notes but not in an analyzable form. Adverse-event rates cannot be reported from this cohort because structured safety information was not captured. Similarly, adherence to medications, self-care measures and splint wear were queried clinically but not recorded as analyzable variables. The investigators explicitly state these omissions and that they limit interpretation of between-group differences.
The study’s sole endpoint was a patient-reported global rating of change at three months, using a four-point scale (much better, better, improved, not improved). For analysis the first three categories were combined as improved and compared with not improved. Improvement was recorded in 47 of 54 chewers (87.0%) and 88 of 94 non-chewers (93.6%), an absolute difference of 6.6 percentage points (95% CI −16.8 to 3.7). Fisher exact p = 0.23 for the unadjusted comparison. The crude odds ratio for improvement among chewers was 0.47 (95% CI 0.15–1.42); after adjustment for age and sex the OR was 0.92 (95% CI 0.26–3.29). Age emerged as the only consistent predictor of non-improvement in adjusted models; chewers were older on average.
Conservative management benefited roughly nine out of ten patients in this cohort. The investigators caution, however, that the study was powered only to detect a large absolute difference (≥16.1 percentage points) and therefore could not exclude a smaller but clinically meaningful disadvantage to chewers given the observed 6.6-point gap and limited events.
Baseline pain intensity was recorded on an 11-point Numeric Rating Scale and categorized for analysis into 0, 1–3 (mild), 4–6 (moderate) and 7–10 (severe). The pattern of pain intensity differed between chewers and non-chewers: among chewers 29.6% reported no pain, 24.1% severe pain and none fell into the mild band, producing a different band shape (p < 0.001) though not a different rank (p = 0.22). Among participants reporting pain, 38 of 40 chewers rated it moderate or severe compared with 52 of 82 non-chewers. The investigators interpret this as a qualitative difference in presentation rather than solely greater severity.
Four covariates were prespecified for adjusted analyses: age, sex, self-reported emotional stress and self-reported tooth grinding. Adjusted estimates relied on a small number of events: thirteen patients in total failed to improve. The authors therefore describe adjusted results as hypothesis-generating. The cohort provided 80% power only to detect an absolute difference of 16.1 percentage points or greater; power for the observed 6.6-point difference was 0.26. These calculations frame the interpretation of a non-significant p-value and a confidence interval that includes clinically important values.
The study authors report several important limitations. They used RDC/TMD rather than the current DC/TMD diagnostic standard; the single-examiner, non-blinded design risks bias in the direction of exaggerating any disadvantage among chewers. Several clinically relevant variables were captured only in routine clinical records and not entered into the research dataset: prescribed doses, dosing schedules, splint design and structured adverse-event reporting. Adherence measures were not recorded in an analyzable way and therefore cannot explain between-group differences. Some continuous measures were stored only as ordered categories in the electronic dataset, preventing computation of change scores for pain and opening. Finally, the sample size allowed detection only of large effects; the study cannot exclude smaller but clinically meaningful disadvantages for habitual chewers. These limitations leave open specific questions documented by the team: whether drug metabolism differences (theoretical interactions with cytochrome P450) affect treatment efficacy in chewers; whether detailed adherence or dosing differences explain any outcome gap; and how outcomes would compare when using DC/TMD diagnostic criteria or blinded assessment.
The report ends on those documented limitations and unanswered questions. The investigators present their adjusted analyses as exploratory and recommend cautious interpretation of a short-term advantage or disadvantage linked to habitual qat chewing, given the dataset and constraints they have recorded.