XSci

Persistent Multisystem Physical Symptoms With Gastrointestinal Manifestations: A Self-Reported Case of Improvement Temporally Associated With Fluoxetine Use

Galal Mohammed Ayad Qaid AlKarabi

Submitted by Galal Mohammed Ayad Qaid AlKarabi

Published November 13, 2026 · Version v1, November 13, 2026 · DOI 10.66977/xsci.2611.0023

Clinical Medicine

Abstract

Abstract

Background: Persistent physical symptoms involving multiple systems can be difficult to characterise, particularly when investigations do not identify a single explanatory condition. A functional or psychiatric diagnosis should not be inferred solely from unrevealing investigations.

Case presentation: A 28-year-old man, who was a medical student during the period described, reported approximately two years of dizziness, blurred vision, fatigue during prolonged sitting, epigastric discomfort, abdominal gas and distressing nocturnal abdominal distension. Symptoms affected concentration and academic functioning and were reported to worsen during periods of academic pressure. Examination was described as unremarkable apart from mild yellowish discoloration confined to the palms. Available laboratory results included haemoglobin 13.0 g/dL, mean corpuscular volume 60 fL, red blood cell count 6.7 × 10⁶/µL, white blood cell count 5,000/µL and ferritin 90 ng/mL. Stool microscopy reportedly identified amoebic cysts, the clinical significance of which was uncertain. Previous empirical treatments, including pantoprazole, antispasmodics, iron supplementation and repeated metronidazole courses, were not associated with sustained benefit. The patient independently formulated a possible somatic symptom disorder explanation; this was not established by an independent psychiatric assessment. He self-initiated paroxetine 20 mg daily for six months, reporting partial improvement and sexual adverse effects (delayed ejaculation and decreased libido). He then switched directly to fluoxetine 20 mg daily without documented medical supervision. Dizziness and blurred vision reportedly resolved within approximately eight weeks, with subsequent improvement in gastrointestinal symptoms. He reported sustained improvement during the first 18 months of daily fluoxetine use; thereafter, he used fluoxetine intermittently.

Conclusion: This report documents a patient's retrospective account of improvement temporally associated with fluoxetine use. It cannot establish a psychiatric diagnosis, treatment efficacy or causation. Marked microcytosis without anaemia and the stool microscopy finding require separate interpretation and, where clinically appropriate, further evaluation.

View PDF

Review reports

xPeerd (AI review)

November 13, 2026

AI Review (xPeerd — DAReview)

Summary

This case report describes a retrospective, self-reported account of a 28-year-old man with approximately two years of dizziness, blurred vision, fatigue, epigastric discomfort, abdominal gas, and nocturnal abdominal distension, with symptoms reportedly worsening during academic pressure. The manuscript’s main strength is its repeated caution against overclaiming: it explicitly states that the report “cannot establish a psychiatric diagnosis, treatment efficacy or causation” and that the patient’s somatic symptom disorder formulation was not independently verified . The paper also appropriately separates the fluoxetine-associated improvement narrative from unresolved findings such as marked microcytosis and stool microscopy showing amoebic cysts .

Its main weaknesses are methodological and editorial rather than conceptual. The evidence base is intrinsically limited by retrospective self-report, absence of independent psychiatric assessment, incomplete gastrointestinal and hematological work-up, lack of standardized outcome measures, and the author–patient dual role, all of which the manuscript itself acknowledges . The central claim is therefore only a temporal association, not a clinically persuasive treatment observation. Overall, the manuscript is cautious and more disciplined than many anecdotal reports, but it still requires substantial revision to improve reporting rigor, ethical clarity, reproducibility, and presentation before it would be suitable for submission.

Potential Major Revisions

  1. Clarify the manuscript’s true contribution and narrow the claim further. The title states “Persistent Multisystem Physical Symptoms With Gastrointestinal Manifestations: A Self-Reported Case of Improvement Temporally Associated With Fluoxetine Use” . This is appropriately cautious, but the manuscript still risks being read as an implicit therapeutic signal because the narrative gives substantial prominence to symptom improvement after switching medication. The Discussion correctly states that “It does not demonstrate that the symptoms were caused by a psychiatric disorder or that fluoxetine caused the subsequent improvement” and that “The case cannot establish an SSRI-specific effect or support fluoxetine as a treatment recommendation for similar patients” . I think this caution should be strengthened structurally, not just verbally.

The paper would be more coherent if framed primarily as a reporting exercise in diagnostic uncertainty and self-attributed recovery rather than a treatment-associated case. As written, the fluoxetine trajectory occupies the narrative center, whereas the deeper scientific issue is uncertainty: “Important elements of the gastrointestinal work-up are not documented” , and “The cause of microcytosis remains undetermined” . The manuscript should explicitly reposition its contribution as an illustration of why retrospective improvement narratives must remain non-causal in the absence of diagnostic confirmation and objective follow-up.

  1. Strengthen methodological transparency regarding chronology and prior work-up. The case history contains important omissions that materially limit interpretability. The manuscript states that “The dates, doses, durations and objective indications for these interventions were not available” for prior treatments including pantoprazole, antispasmodics, iron supplementation, and metronidazole . It also states, “The exact interval between the last paroxetine dose and the first fluoxetine dose, and any cross-tapering schedule, were not documented” . These are not minor details; they directly affect whether the observed change can even be temporally interpreted with minimal confidence.

The manuscript should add a more complete chronology table including approximate dates, symptom phases, consultations, investigations, all medication exposures, treatment durations, adverse effects, and any contextual stressors. If exact dates are unavailable, that uncertainty should be shown explicitly rather than embedded narratively. At present, the timeline is helpful but still too sparse to support reproducible reading of the case trajectory .

  1. Address the incompleteness of diagnostic assessment more directly. The manuscript commendably avoids claiming confirmed SSD and explicitly notes that “No structured psychiatric interview, documented assessment of DSM-5-TR criterion B, or validated symptom measure such as the PHQ-15 was available” . It also notes missing medical investigations, including whether testing for “Helicobacter pylori, coeliac disease, thyroid dysfunction, vitamin B12 deficiency, glucose abnormalities or endoscopic evaluation was performed” .

However, these omissions are so central that they should be elevated from supporting limitation to principal interpretive barrier. The manuscript currently reads as though SSD, gut–brain interaction disorders, functional dyspepsia, bloating syndromes, hematologic explanations, and infectious explanations were all considered but unresolved. That is acceptable descriptively, but the paper should more clearly explain why this unresolved differential diagnosis substantially weakens the scientific value of attributing symptomatic improvement to one pharmacologic event. Without this, the report risks reproducing the logic it seeks to resist: unexplained symptoms followed by medication change followed by presumed explanatory narrative.

  1. Improve the handling of the author–patient dual role and possible bias. The manuscript openly states: “The author is also the patient described in this report” and later acknowledges that this “introduces potential recall, selection and reporting bias” . This transparency is a strength. Still, the paper should do more to show how bias was mitigated. For example, were any laboratory reports, prescription packages, contemporaneous notes, academic records, or other dated materials used to reconstruct the chronology? The statement that “not all details could be corroborated by contemporaneous clinical documentation” implies that some may have been, but the manuscript does not specify which ones.

A dedicated subsection should distinguish directly documented facts from recalled impressions. For instance, lab values appear more concrete, while symptom timing and functional gains appear more subjective. This separation would materially improve credibility.

  1. Expand the ethical and editorial compliance discussion. The Declarations state: “No institutional research ethics committee approval was sought or obtained for this case report, and no formal waiver or exemption was obtained” and that “The journal's editorial office must be consulted before submission” . This is an unusually candid disclosure, but it also signals a potential submission barrier. For many journals, case reports involving self-report and self-authorship may still require formal ethics clarification, even if not full committee review.

The manuscript should not leave this as an open procedural question. Before submission, the author should obtain either an ethics determination, a waiver, or explicit documentation that such approval is not required under the relevant institutional or journal policy, and then revise the ethics statement accordingly. In its current form, the paper appears procedurally unfinished.

  1. Deepen the analysis of alternative explanations for improvement. The Discussion lists plausible alternatives: “fluctuating natural history of symptoms, changes in academic stress, delayed effects of prior treatment, regression to the mean, expectancy effects, and unmeasured changes in health behaviours or other treatments” . This is good practice. Yet these possibilities are mentioned only briefly. They should be integrated more analytically into the case interpretation.

For example, the manuscript reports that symptoms “were reported to worsen during periods of academic pressure” and that recurrences later occurred “during accumulated life stress” . This pattern suggests a potential stress-linked course independent of any drug-specific effect. Because this is one of the strongest alternative interpretations available from the manuscript’s own evidence, it deserves more direct analysis in the Discussion.

  1. Fraud/paper-mill screening assessment. I do not see strong paper-mill indicators such as incoherent jargon, fabricated citation style, impossible claims, or internally contradictory methodology. In fact, the manuscript is unusually explicit about its limitations and non-causal claims . That said, there are integrity-sensitive features: self-authorship, retrospective reconstruction, absence of independent assessment, and missing documentation for key clinical events . These do not indicate fraud by themselves, but they do reduce evidentiary reliability. The paper fits a recognizable clinical case-report paradigm, though presently at the weak end of that paradigm because its principal observations are largely unverifiable and subjective.

Potential Minor Revisions

  1. Typographic, grammatical, and style issues
Location Quoted text Issue Suggested revision
Declarations “Conflicts of interest: The author confirm no conflict of interest” Subject–verb agreement; awkward phrasing “Conflicts of interest: The author confirms that there are no conflicts of interest.”
Author line / front matter “ORCID:0009-0009-9347-1824@orcid” Nonstandard ORCID formatting Use standard format, e.g. “ORCID: 0009-0009-9347-1824”
Keywords “persistent physical symptoms; gastrointestinal symptoms; fluoxetine; paroxetine; microcytosis; case report” Acceptable, but “persistent physical symptoms” and “gastrointestinal symptoms” may be too broad for indexing Consider more precise terms such as “somatic symptoms,” “functional gastrointestinal symptoms,” or “self-reported case” depending on target journal
Patient perspective “I considered the improvement sustained while I took fluoxetine daily for the first 18 months” Grammatically fine, but repetitive with prior prose Shorten to avoid duplication with Results/Follow-up
Declarations “Generative AI: ChatGPT (OpenAI) was used to assist with language editing, only” Comma placement is awkward “Generative AI: ChatGPT (OpenAI) was used only for language editing.”
  1. Formatting and structure issues
Area Issue Comment
Pagination Search results repeatedly show “Page 1,” suggesting page-number extraction is unreliable or the PDF pagination may not be properly embedded Ensure the submitted PDF has stable pagination so reviewers can cite accurately
Timeline presentation The timeline is useful but appears fragmented in the extracted text Reformat into a cleaner journal-style table with dates/intervals, events, symptoms, and treatments
Section balance Discussion and limitations are relatively strong, but diagnostic assessment and objective evidence remain brief compared with narrative interpretation Expand objective clinical documentation if available
Declaration completeness Ethics language signals uncertainty rather than compliance Replace with finalized editorially compliant wording before submission
Title emphasis The current title foregrounds fluoxetine improvement Consider a title that foregrounds diagnostic uncertainty and retrospective self-report to reduce therapeutic implication
  1. AI content analysis for a post-2021 work

Estimated percentage of AI-generated content: Low to moderate likelihood of AI-assisted editing at sentence level, but not strong evidence of wholly AI-generated scientific content.

Rationale: The manuscript itself discloses, “Generative AI: ChatGPT (OpenAI) was used to assist with language editing, only” . The prose is unusually consistent, clean, and caution-heavy, which is compatible with professional editing or AI-assisted language smoothing. However, the manuscript also contains individualized clinical specifics, self-limiting statements, and coherent internal cross-checking that make it look more like a human-authored text polished by AI than an AI-fabricated manuscript.

Highlighted AI-detected sections: The strongest AI-assisted signature appears in standardized cautionary prose such as “The temporal association between fluoxetine use and reported improvement should be interpreted cautiously” and the subsequent balanced list of alternative explanations . Similarly, the phrasing in the limitations and key learning points is polished, formulaic, and editorially normalized . These sections may well have been language-edited with AI, consistent with the disclosure.

Assessed epistemic impact: The disclosed AI assistance does not appear to be the main problem. The epistemic weakness arises from retrospective self-report, missing verification, incomplete work-up, and nonstandardized outcomes, not from AI-edited language. If AI was used only for language editing, as disclosed, its impact on the manuscript’s validity is minor compared with the methodological limitations already acknowledged in the text .

Recommendations

  1. Reframe the manuscript so that its principal contribution is diagnostic and interpretive caution, not medication-associated improvement.
  2. Add a more complete chronology table with approximate dates, investigations, treatments, symptoms, stress periods, and adverse effects.
  3. Separate documented facts from recalled facts in the case presentation.
  4. Expand the Discussion of alternative explanations, especially stress-related fluctuation and regression to the mean.
  5. Strengthen the differential diagnosis discussion by explaining how missing gastrointestinal and psychiatric assessments limit interpretation.
  6. Finalize ethics and editorial compliance before submission rather than noting that the journal must be consulted.
  7. Revise declarations and copyediting details, especially the conflict-of-interest sentence and AI disclosure phrasing.
  8. If possible, append supporting source materials such as laboratory reports or contemporaneous records to improve verifiability.
  9. Consider whether the manuscript would be stronger as a brief reflective case note or hypothesis-generating correspondence rather than a full clinical case report.
  10. Maintain the manuscript’s current restraint regarding causation; this is one of its strongest features and should not be diluted in revision.

Reviewer3 (AI review)

November 13, 2026

AI Review (reviewer3.com)

Overview

Multisystem Symptom Case Narrative

Contribution

The author documents a self-reported case of persistent multisystem physical and gastrointestinal symptoms that temporally resolved following self-directed fluoxetine treatment.

Diagnostic Limitations in Somatic Symptoms

Contribution

The paper illustrates critical diagnostic caveats, demonstrating that unrevealing medical investigations and retrospective self-reports are insufficient to establish a somatic symptom disorder or infer pharmacological causality.

Comments (2)

1. Unresolved pre-submission drafting text and missing self-reported details — Critical

information. The records supplied for this report do not document whether testing for Helicobacter pylori, coeliac disease, thyroid dysfunction, vitamin B12 deficiency, glucose abnormalities or endoscopic evaluation was performed. These tests are not assumed to have been negative.

Although the author is the patient described in this report, the manuscript is written in the third person as if reviewing an incomplete external file (stating that details are missing from the records supplied for this report and failing to specify the dose of fluoxetine tablets the author himself took intermittently), and the Declarations section retains pre-submission instructions. Because this is a self-reported retrospective account, please clarify what you recall versus what was documented in medical records, provide the approximate duration of follow-up after the initial 18-month period, and finalize all pre-submission placeholder text.

2. Missing laboratory reference ranges for haematological and biochemical values — Major

count, is approximately 9.0. The haemoglobin value does not indicate anaemia in the context provided, while the MCV indicates marked microcytosis.

Haemoglobin (13.0 g/dL) is interpreted as not indicating anaemia, and C-reactive protein, liver function, and kidney function are described simply as normal without providing local laboratory reference intervals or numerical values. Because a haemoglobin of 13.0 g/dL is borderline for an adult male (particularly at high altitude, such as in Ibb, Yemen, where haemoglobin reference thresholds are higher), please include the specific laboratory reference ranges and numerical results where available to support the characterization of microcytosis without anaemia.

XSci (AI review)

November 13, 2026

Summary

The authors present a single autobiographical case report: a 28-year-old medical student with approximately two years of dizziness, blurred vision, fatigue on prolonged sitting, epigastric discomfort, abdominal gas and nocturnal distension, whose investigations showed haemoglobin 13.0 g/dL, MCV 60 fL, RBC 6.7 × 10⁶/μL, ferritin 90 ng/mL and amoebic cysts on stool microscopy. After pantoprazole, antispasmodics, iron and repeated metronidazole gave no sustained benefit, the patient self-initiated paroxetine 20 mg daily for six months with partial improvement and sexual adverse effects, then switched directly to fluoxetine 20 mg daily. Dizziness and blurred vision reportedly resolved within approximately eight weeks, with later gastrointestinal improvement sustained over 18 months of daily use. The report states throughout that it cannot establish a diagnosis, efficacy or causation, and refers the microcytosis and the stool finding for separate evaluation.

Contribution and Significance

The contribution is methodological rather than clinical. The manuscript's real subject is how not to reason about a persistent-symptom presentation: it declines to convert an unrevealing work-up into somatic symptom disorder, naming the specific deficiency (DSM-5-TR criterion B was never assessed, and no structured interview or PHQ-15 exists); it declines to convert a Mentzer index of 9.0 into thalassaemia trait, naming the missing confirmatory tests; and it declines to let amoebic cysts explain a two-year multisystem syndrome. That restraint is genuinely uncommon in this literature and the authors execute it consistently.

Relative to the closest prior work, however, nothing new is established. The SSRI question in these presentations is already addressed by randomised and synthesised evidence the authors themselves cite accurately and deflationarily: Kleinstäuber et al. (Cochrane 2014) [7], Talley et al. (Gastroenterology 2015) [9], and Ford et al. [8,10]. A single uncontrolled, undocumented, retrospectively self-reported temporal association cannot move that evidence base, and the manuscript agrees. The difficulty is that the five Key learning points are all general principles that hold without this case; none is derived from it. The two case-specific findings are flagged and then explicitly deferred for evaluation that was never performed, so neither teaches anything either.

Strengths

  • The refusal to promote the patient's self-formulation to a diagnosis is specific and correctly reasoned (Diagnostic assessment, fourth paragraph; Discussion, second paragraph), not a generic disclaimer.
  • The haematological arithmetic is correct and appropriately bounded: MCV/RBC = 60/6.7 ≈ 9.0, explicitly labelled a screening aid with electrophoresis and HPLC named as the missing confirmatory tests [4,5].
  • Incidental findings are deliberately kept separate rather than welded into one unifying diagnosis (Conclusion; Key learning points 4–5).
  • The dual patient-author role is declared in the body text and its recall, selection and reporting biases discussed (Discussion, seventh paragraph; Declarations), not buried.
  • The CARE architecture is complete, including a Timeline table and a genuine first-person Patient perspective.

Recommendation

Reject for the Journal of Medical Case Reports. For this venue the advance is too small and too unclear to matter: the manuscript disclaims every assertion it could make, and what remains is an unverifiable anecdote plus five textbook generalities that are stateable without the case. A secondary consideration is that the report is too incomplete to assess — there is no corroborating documentation, no dates, no independent assessment and no outcome measure. I note that the gaps are largely remediable, and Concern 3 sets out what a stronger, different manuscript would contain.

Major Concerns

  1. No case-specific teaching point survives the manuscript's own disclaimers (Abstract Conclusion; Discussion, first and fourth paragraphs; Conclusion; Key learning points). The Abstract states the report "cannot establish a psychiatric diagnosis, treatment efficacy or causation", and the Discussion adds that it "cannot establish an SSRI-specific effect or support fluoxetine as a treatment recommendation". I agree with every one of these statements, which is precisely the problem: taken together they leave no claim standing. Each of the five learning points — that SSD requires positive features, that an N-of-1 cannot establish causation, that self-switching is not a strategy, that microcytosis without anaemia needs evaluation, that cysts do not establish invasive amoebiasis — is true independently of this patient and could be written without having met him. A case report must earn its place by what the case shows; the authors should identify the one thing this case shows that a textbook paragraph does not, and build the manuscript on it.

  2. The two most plausible explanations for the reported improvement are never examined (Discussion, fourth paragraph). The alternatives offered are fluctuating natural history, academic stress, delayed effects of prior treatment, regression to the mean, expectancy effects and unmeasured changes in health behaviours. Absent from that list is an undiagnosed depressive or anxiety disorder responding to an SSRI — yet the Diagnostic assessment states plainly that "Depression and anxiety were not formally assessed in the reported record", symptoms were reported to worsen under academic pressure, and the sequence of partial response to one SSRI followed by response to another is the expected shape of exactly that. This is the leading hypothesis and it goes unnamed.

    Also absent is the differential pharmacology of the two drugs. Paroxetine has the greatest antimuscarinic affinity of the SSRIs, and the symptoms reported to resolve first after the switch — blurred vision and dizziness, at approximately eight weeks — are among those paroxetine is most likely to sustain. Whether or not that explanation is correct, it is the one mechanistic account the case is actually positioned to raise, and engaging with it would give the report content it currently lacks.

  3. The missing information is largely retrievable by the author, who is the patient (Patient information; Clinical findings; Diagnostic assessment, fifth paragraph; Previous treatments; Follow-up). The manuscript repeatedly uses third-party framing — "the material available for this report", "the records supplied for this report", "the available account does not specify" — which is incoherent when the Declarations confirm the author and the patient are the same person. An author-patient can state whether fundoscopy, H. pylori testing, endoscopy, TSH or B12 were done rather than listing them as undocumented; can supply dates and the timing of the ferritin relative to the iron supplementation; can obtain the RDW and platelet count from the same automated blood count; and can obtain the haemoglobin electrophoresis the manuscript itself identifies as missing. Presenting remediable gaps as evidentiary limitations understates what the report could contain and overstates how little could be known.

Minor Comments

  1. Haemoglobin 13.0 g/dL is exactly the WHO lower limit for adult men, so "marked microcytosis without anaemia" is borderline rather than clear-cut and should be qualified wherever it appears.
  2. RDW and platelet count are not reported although both appear on the same automated count, and RDW discriminates thalassaemia trait from iron deficiency better than the Mentzer index.
  3. The ferritin of 90 ng/mL is reported without stating its timing relative to the iron supplementation described under Previous treatments, which affects its interpretation.
  4. The only abnormal physical sign reported, "mild yellowish discoloration limited to the palms" (Clinical findings), is never revisited in the Discussion; carotenoderma and hypothyroidism should be addressed or the finding dropped.
  5. Reference [1] contains an author error: the CARE author is Altman DG, not "Altman UG" (Gagnier JJ, Kienle G, Altman DG, Moher D, Sox H, Riley D; CARE Group. J Med Case Rep. 2013;7:223).
  6. Reference [1] is a reporting checklist and does not support the Introduction's statement about choosing diagnostic terminology; a CARE-adherence statement belongs in the Declarations instead.
  7. Metronidazole is a tissue amoebicide and does not reliably clear luminal cyst passage, which is directly relevant to the repeated courses described under Previous treatments.
  8. No calendar dates appear anywhere; year-level anchoring would let readers judge the duration of the intermittent-use period.
  9. The paternal-uncle history of schizophrenia (Patient information) is introduced, set aside, and never used; give its relevance or remove it.
  10. Key learning point 3 would be more useful if it named the specific hazard — fluoxetine's long half-life and CYP2D6 inhibition acting on a recently discontinued paroxetine — rather than stating only that self-switching is not recommended.
  11. The Ethics declaration leaves the ethics determination unresolved and consent is self-consent by the author-patient; both should be settled with the editorial office before submission rather than deferred within the manuscript.
  12. The statement that causation cannot be established appears in the Abstract, twice in the Discussion, in the Limitations and in the Conclusion; once would suffice and would free space for substantive content.