Source Case: 60F with Coma, E Coli Sepsis and Upper Motor Neuron Signs with Albumino-Cytological Dissociation in CSF HIPAA de-identified open online patient record, December 2016. BMJ Elective Student, informed patient consent obtained.
Companion document to: VibeRounds-CaseAnalysis-60F-EColi-Sepsis.md
This document applies Module 9 Step 9.8 (CARE-field export) and Modules 2, 3 & 11 (Advocate Debrief) to the same case.
CARE Guideline Reference: Gagnier JJ et al., “The CARE guidelines: Consensus-based clinical case report guideline development,” Journal of Clinical Epidemiology, 2014. The CARE checklist requires: title, abstract, introduction, patient information, clinical findings, timeline, diagnostic assessment, therapeutic interventions, follow-up and outcomes, discussion, patient perspective, and informed consent statement.
Delayed diagnosis in a 60-year-old woman with an 18-year history of cervical myelopathy presenting with coma, MDR Escherichia coli sepsis, pancytopenia, stony splenomegaly, and albumino-cytological dissociation in cerebrospinal fluid: a case of missed systemic disease.
Background: Cervical myelopathy is a common diagnosis in older adults that can anchor clinical reasoning for years and obscure a concurrent or underlying systemic disease. We report a case where an 18-year working diagnosis of degenerative cervical myelopathy prevented timely recognition of a multi-system disease characterised by pancytopenia, stony splenomegaly, hepatomegaly, ascites, and albumino-cytological dissociation in cerebrospinal fluid (CSF).
Case Summary: A 60-year-old woman with an 18-year history of cervical spinal disease, managed conservatively with baclofen, presented with a 4-month acceleration of neurological symptoms, fever, progressive anaemia refractory to transfusion, and eventually coma, type 2 respiratory failure, and MDR E. coli bloodstream infection. Investigations revealed pancytopenia with normocytic anaemia, stony hard splenomegaly, soft hepatomegaly, ascites, CSF albumino-cytological dissociation, and bone marrow erythroid hyperplasia. CT brain was normal; cervical MRI was near-normal. Despite treatment for sepsis and supportive care including mechanical ventilation, the patient died. No unifying systemic diagnosis was established before death.
Key Learning Points: (1) Stony splenomegaly, pancytopenia, and CSF albumino-cytological dissociation together constitute a triad that demands systemic investigation independent of the neurological working diagnosis. (2) CSF cell count is unreliable in pancytopenic patients; near-normal cells do not exclude active meningeal infection in this context. (3) Baclofen toxicity is a pharmacological differential for coma, hypothermia, and type 2 respiratory failure in the setting of hepatic or renal impairment, and should not be overlooked when these features co-exist. (4) Anchoring bias operating over 18 years can survive multiple acute deteriorations and multiple healthcare encounters without being challenged.
Cervical myelopathy is among the most common causes of progressive spinal cord dysfunction in adults over 50 and is frequently managed conservatively for extended periods. Its chronicity makes it a powerful anchoring diagnosis — clinicians encountering a patient with a documented multi-year history of cervical spine disease are cognitively predisposed to interpret new or worsening symptoms within that existing frame.
This case illustrates the clinical and cognitive consequences of that anchoring over nearly two decades. The patient described here had genuine cervical spinal disease; the diagnosis was not wrong. What was missed was the concurrent or underlying systemic process that emerged — or accelerated — over the final four months of her life, producing a clinical picture that cervical myelopathy alone cannot explain: stony hard splenomegaly, pancytopenia, hepatomegaly with ascites, albumino-cytological dissociation in CSF, refractory anaemia, and ultimately coma and death from MDR gram-negative sepsis in what was almost certainly an immunocompromised host.
The educational contribution of this case is not the rarity of the underlying condition — it may have been disseminated tuberculosis, haematological malignancy, or autoimmune multi-system disease, all of which are well-described — but rather the anatomy of the diagnostic failure: which clinical signals were present, when they appeared, what investigations were available, and at what point the cognitive frame should have been reset. We report this case to document that anatomy and to offer a structured framework for recognising similar missed pivots in future practice.
| Field | Detail |
|---|---|
| Age | 60 years |
| Sex | Female |
| Relevant background | Coal-mining area resident for most of life; relocated to urban area 18 years prior |
| Dietary history | Non-vegetarian (white meat primarily); no alcohol or tobacco in family |
| Cooking fuel | Chulha (biomass/wood fire) until 20 years prior; urban fuel thereafter |
| Occupation | Homemaker |
| Social support | Family-managed care throughout; son-in-law/family as primary advocates |
| Comorbidities documented | None — no diabetes mellitus, no hypertension |
| Prior diagnoses | Cervical spinal disease (myelopathy, likely degenerative or compressive) 18 years prior; H. pylori-positive gastric ulcers (identified during current admission) |
| Patient perspective | Documented through family narrative (see Part B); patient was non-communicative during terminal admission |
| Parameter | Finding | Clinical Significance |
|---|---|---|
| Temperature | 95°F (35°C) | Hypothermia — septic shock vs. autonomic failure vs. drug effect |
| Blood pressure | 159/70 mmHg | Relative hypertension — may reflect pain, sympathetic activation, or pre-existing tendency |
| Respiratory rate | 33/min on ventilator; 16/min spontaneous | Ventilator-dependent; laboured effort off ventilator |
| Consciousness | Non-responsive, comatose | GCS not recorded in available narrative |
| Pallor | ++ | Consistent with severe anaemia (Hb 5.6–5.7 g/dL) |
| Oedema | + pitting | Hypoalbuminaemia, cardiac, or hepatic cause |
| Spleen | Stony hard, enlarged | Infiltrative process (TB, lymphoma, myeloproliferative disease) — NOT consistent with simple portal hypertension |
| Liver | Soft, enlarged | Hepatomegaly; texture suggests reactive or infiltrative rather than cirrhotic |
| Ascites | Present | Hepatic, peritoneal, or lymphatic origin |
| Skin | Sacral pressure sores, 2nd degree | Prolonged immobility, sensory loss, nutritional depletion |
| Neurological | UMN signs (spasticity, weakness); altered sensorium | Myelopathy + encephalopathy |
| Rectal | History of rectal fistula; melaenic stool | Active or recent GI haemorrhage |
| Investigation | Result | Interpretation |
|---|---|---|
| Haemoglobin | 5.6 g/dL → 5.7 g/dL after 2 units packed red cells | Refractory to transfusion — implies active loss or haemolysis |
| Full blood count | Pancytopenia (all three cell lines reduced) | Bone marrow failure, infiltration, or peripheral destruction |
| Differential count | PMN predominance | Suggests bacterial infection as acute trigger; does not exclude underlying lymphopenia |
| Renal function | Relatively preserved | Hypokalemia documented |
| Liver function | Not documented in available records — a significant gap | |
| CSF | High protein, normal glucose, ~5 cells | Albumino-cytological dissociation |
| CSF ADA | Elevated (referenced in clinical discussion) | Raises TB meningitis as possibility |
| Bone marrow biopsy | Erythroid hyperplasia | Reactive pattern — peripheral red cell destruction or haemorrhage; does not exclude infiltrative disease |
| Blood culture | MDR E. coli | Gram-negative sepsis; implies immunosuppression or GI/urinary source |
| CT brain | Normal | No mass lesion, infarct, or haemorrhage |
| MRI cervical spine | Near-normal (done at external centre) | Limits structural myelopathy as the primary driver of acute deterioration |
| H. pylori | Positive | Gastric ulcers confirmed |
| ABG (serial) | Rising pCO₂, falling pH, normal bicarbonate | Type 2 respiratory failure with respiratory acidosis |
| Missing — not obtained | LFT, peripheral blood smear morphology, ANA/ANCA/anti-dsDNA, serum protein electrophoresis, TB-PCR from CSF and bone marrow, AFB culture from bone marrow, HRCT chest, autoimmune panel | These investigations could have established or excluded the principal differential diagnoses |
18 years prior
│ Neck symptoms begin: electric sensations in limbs, nausea, vomiting, vertigo
│ Cervical myelopathy diagnosed; surgery advised but declined by family
│ Baclofen prescribed PRN (taken 2–3 times/week for ~17 years)
│
4 months prior to terminal admission
│ Symptoms accelerate: neck pain, electric sensations (whole body), headache,
│ breathlessness, loss of sensation in little/ring fingers, unable to walk
│ Baclofen escalated to daily then twice-daily (patient-initiated)
│
~4 months prior
│ First hospital admission: surgery planned → cancelled (pancytopenia identified)
│ H. pylori ulcers found simultaneously
│ Referred back with instruction to normalise blood counts
│
~3.5 months prior
│ Second hospital admission (ICU): blood transfusion + albumin
│ Brief improvement; nasogastric tube placed for feeding
│ Discharged home with NGT in situ
│
~3 months prior (home)
│ Day 8 at home: diarrhoea (watery, blackish, 4–5 episodes/day)
│ Self-resolving over 3 days without treatment
│ Pomegranate juice stopped (attributed to black stool colour)
│
~2.5 months prior
│ Breathlessness worsens; chest and back pain; nausea and vertigo return
│ Hb measured at 5.6 g/dL
│ Two units of blood given on GP advice → Hb rises only to 5.7 g/dL
│ One further unit given next day
│
~10 days prior to death
│ Severe breathlessness
│ Loss of consciousness
│ Admitted to ICU; ventilated
│ MDR E. coli sepsis confirmed in blood
│ CSF: albumino-cytological dissociation
│ Bone marrow: erythroid hyperplasia
│ CT brain: normal; cervical MRI: near-normal
│
Terminal admission
│ Hypothermia (95°F); BP 159/70; RR 33 on ventilator
│ Non-responsive; pancytopenia; stony splenomegaly; soft hepatomegaly; ascites
│ Sacral pressure sores 2nd degree
│ Type 2 respiratory failure (serial ABG: rising pCO₂, respiratory acidosis)
│
Death
No unifying systemic diagnosis established before death
| Diagnosis | Supporting Evidence | Against | Investigation Required |
|---|---|---|---|
| Disseminated tuberculosis | Coal-mine history, elevated CSF ADA, pancytopenia, stony splenomegaly, hepatomegaly, ascites, myelopathy, CSF dissociation | No HRCT chest done; MRI spine near-normal | TB-PCR CSF, AFB smear/culture bone marrow, HRCT chest |
| Haematological malignancy (lymphoma / MPD) | Stony splenomegaly, pancytopenia, hepatomegaly, CSF dissociation (carcinomatous meningitis), refractory anaemia | Bone marrow showed erythroid hyperplasia not infiltration (though single biopsy is not definitive) | Flow cytometry on bone marrow, repeat LP with CSF cytology, serum LDH and protein electrophoresis |
| Autoimmune multi-system disease (SLE/Sjögren’s) | Myelopathy, pancytopenia, serositis (ascites), CSF protein elevation | No prior joint symptoms; no rash documented | ANA, anti-dsDNA, anti-Ro/La, C3/C4, Coombs test |
| Baclofen toxicity (contributing) | Hypothermia, coma, Type 2 respiratory failure, escalating dose in context of possible hepatic impairment | Difficult to confirm without serum level | Serum baclofen level (where available); clinical: careful dose reduction with serial neurological observation |
| Hepatic myelopathy | Soft hepatomegaly, ascites, splenomegaly, possible portosystemic shunting | No cirrhotic texture, no varices documented | LFT, liver biopsy if stable enough, portal imaging |
Does a single systemic disease explain: cervical myelopathy (or myelitis), stony splenomegaly, pancytopenia, hepatomegaly, ascites, CSF albumino-cytological dissociation, refractory anaemia, and immunosuppression permitting MDR gram-negative sepsis?
Disseminated tuberculosis remains the most likely unifying diagnosis given the epidemiological and clinical profile, though haematological malignancy with secondary immunosuppression cannot be excluded. These diagnoses were diagnostically accessible before death and should have been pursued from the point at which surgery was cancelled due to pancytopenia — a clinically visible turning point four months before death.
| Intervention | Timing | Rationale | Outcome |
|---|---|---|---|
| Baclofen (PRN → daily → twice-daily) | Ongoing 18 years; escalated 4 months prior | Spasticity management | Symptomatic; escalation may have contributed to CNS and respiratory depression |
| Blood transfusion (multiple units) | During and after ICU admissions | Severe anaemia (Hb 5.6–5.7) | Minimally effective; Hb barely rose after 3 units — active destruction/loss |
| Albumin infusion | First ICU admission | Hypoalbuminaemia | Temporary improvement |
| Nasogastric feeding | Post-first ICU discharge | Dysphagia/enteral nutrition | Used for 5 days at home; diarrhoea on day 8 self-resolved |
| Broad-spectrum antibiotics (MDR E. coli) | Terminal admission | Targeted sepsis treatment | Specific antibiotic regimen not documented; outcome fatal |
| Mechanical ventilation | Terminal admission | Type 2 respiratory failure | Patient ventilator-dependent; could not sustain RR >16 spontaneously |
| Pressure sore dressing | Home and hospital | 2nd degree sacral sores | Reportedly improving before re-deterioration |
| Potassium replacement | Terminal admission | Hypokalemia | Not documented as resolved |
The patient did not survive to any follow-up interval. She died during the terminal ICU admission. No post-mortem examination is documented in the available record.
The absence of autopsy represents the final diagnostic gap. Post-mortem examination of the spleen, bone marrow, liver, spinal cord, and meninges would have established the unifying diagnosis with certainty and provided the family with an explanation for the clinical course. In resource-constrained settings where autopsy is not routine, this gap is common and represents a systematic loss of learning that accumulates at the registry level.
On diagnostic anchoring: The 18-year history of cervical myelopathy was the most powerful cognitive anchor in this case. Every acute encounter re-confirmed the existing frame rather than questioning it. The decisive moment — the point at which the frame should have been reset — was the pre-operative discovery of pancytopenia four months before death. Pancytopenia in a patient with myelopathy and stony splenomegaly is not a reason to delay surgery until “blood counts normalise.” It is a reason to stop and ask: what is causing this triad?
On the stony spleen: The texture of the spleen is diagnostically specific in a way that size alone is not. Stony hardness implies infiltration — by granulomas (TB, sarcoid), by malignant cells (lymphoma, leukaemia, myeloproliferative disease), or by fibrosis. None of these causes simple portal hypertension of the kind that produces the soft, enlarged spleen of chronic liver disease. The stony consistency was documented and noted in clinical discussion but was never followed to its logical investigative conclusion.
On CSF interpretation in pancytopenia: The near-normal cell count in CSF (5 cells) was interpreted as excluding active meningitis. This interpretation is unsafe in a pancytopenic patient. When the peripheral white cell count is suppressed, the CSF cell response to meningeal infection is also suppressed — a patient with TB meningitis or carcinomatous meningitis and a white count near zero may present with very few CSF cells despite high protein and high CSF ADA. The protein-cell dissociation in this context may therefore represent active infection or malignant meningeal involvement behind a pancytopenic mask.
On refractory anaemia: Haemoglobin rising only 0.1 g/dL after two units of packed red cells is not a transfusion failure — it is a clinical signal. Active red cell destruction (haemolysis, autoimmune or microangiopathic) or active haemorrhage consuming red cells faster than they are replaced is the only explanation for this pattern. Neither was documented as having been investigated.
On Baclofen: Baclofen toxicity produces a clinical syndrome that in this case is indistinguishable from the terminal presentation: coma, hypothermia, respiratory failure. Baclofen is renally cleared; accumulation occurs with renal or hepatic impairment, and dose toxicity is documented at standard doses when clearance is reduced. The patient self-escalated to twice-daily dosing four months prior. In the absence of renal function testing results at that dose and in the context of hepatic disease, this pharmacological contribution to the terminal presentation should have been explicitly evaluated.
Three accessible investigations, had they been obtained at the point of pre-operative pancytopenia discovery, may have changed the diagnostic trajectory:
All three were available at the level of care this patient was receiving. None are documented as having been obtained. This is the learning contribution of this case.
The following is reconstructed from the family narrative as published in the original open patient record. It represents the family-as-advocate perspective, not a direct patient statement (the patient was non-communicative during the terminal admission).
“18 years ago my mother-in-law had been admitted to hospital for a problem in the neck. She was having a feeling like current in hand and legs, nausea, vomiting, and vertigo… Since it was dangerous we had opted for no surgery and tried to manage on medicine only.”
The family navigated multiple hospital admissions across months, managed nasogastric feeding at home, arranged home dressing of pressure sores, made independent decisions about medication (stopping pomegranate juice, continuing prescribed medications), arranged repeat blood transfusions on GP advice, and continued to advocate for investigation and treatment until the patient’s death. They did this without a clear diagnosis, without understanding why blood counts were not improving, and without ever being told what systemic disease — if any — might be driving the deterioration.
The family’s account reveals several moments where earlier escalation could have been prompted with clearer guidance:
This perspective is addressed fully in Part B.
This case is published as a HIPAA de-identified open online patient record. Informed consent was obtained by a BMJ Elective Student in December 2016 using a consent form publicly available at the source blog. The case is reproduced here for educational purposes with reference to the original public record.
Framework reference: VibeRounds Modules 2 (Patient-Advocate Case Documentation), 3 (Extended Advocate Monitoring), and 11 (Patient Education Query Intelligence), applied retrospectively to this case to reconstruct what the family advocate needed — and what they were not given — at each key clinical moment.
An advocate debrief is a structured retrospective analysis of how a non-medical caregiver navigated a complex clinical case — what information they had, what they lacked, what decisions they made under uncertainty, and what guidance could have changed the outcome at each inflection point. It is not a critique of the family. It is a clinical systems analysis of what the healthcare encounters failed to provide.
The patient’s daughter-in-law (and family collectively) served as:
This level of caregiver responsibility, in the absence of structured advocate education, places the family in a position of carrying clinical risk without clinical knowledge.
What happened: The family declined surgery for cervical myelopathy after being told of “huge complications.” Baclofen was prescribed PRN.
What the advocate understood: Surgery is dangerous. Medicine (Baclofen) will manage the symptoms.
What the advocate needed to know:
What was missing: A structured follow-up plan, a red-flag list, and an escalation threshold. Without these, the family had no framework for the 18 years that followed except “take Baclofen when needed.”
Module 11 Step 11.5 (Red Flag Teaching) applied retroactively:
What the advocate should have been told to watch for:
| Red flag | What it means | Action |
|---|---|---|
| Needing Baclofen more than 3× per week consistently | Condition worsening; medicine not enough | See doctor for review |
| New difficulty breathing | Possible involvement of breathing muscles | Emergency — go to hospital |
| Loss of bladder or bowel control | Significant cord deterioration | Urgent medical review |
| Weakness progressing from legs to arms | Higher cord involvement | Urgent review |
| Any fever lasting more than 3 days | Possible infection; immune risk | Doctor review within 24 hours |
What happened: The patient self-escalated Baclofen from PRN to daily to twice-daily over 4 months as symptoms worsened.
What the advocate understood: She needs more medicine because the symptoms are worse.
What the advocate needed to know:
What the advocate should have been asked to watch for once Baclofen was increased:
The missed signal: The family did not know that these signs — which were all present in the terminal presentation — could have been present intermittently for months as Baclofen accumulated.
What happened: Surgery was cancelled. The family was told to “go home until blood cells become normal.”
What the advocate understood: There is something wrong with the blood. Wait for it to improve.
What the advocate needed to know:
Module 2 Step 2.5 (Data Completeness Audit) applied retroactively:
What was missing from the advocate’s information at this point:
The question the advocate needed to ask but did not know to ask: “If blood counts don’t go up, what are we checking for?”
What happened: The patient was discharged home with an NGT for feeding. A home-visiting nurse dressed the sacral sores.
What the advocate understood: Give food by tube. Continue dressings. The patient is recovering.
What the advocate needed:
Module 2 Step 2.7 (Advocate Handover Brief) — what should have been written:
“Your family member has a serious spinal condition that is worsening. She is being fed through a tube because she cannot safely swallow. She has sores on her back from lying still — these must be dressed daily and she should be turned every 2 hours. Her blood count is low — this means she is at high risk of infection. Watch for: fever above 38°C or below 36°C, any change in level of consciousness, difficulty breathing that is worse than before, blood in stool or very dark black stool, any new wound or swelling. If any of these happen, do not wait — go to the emergency department immediately.”
What was actually given: Not documented. Based on the narrative, the family had no clear escalation threshold — they waited 15 days after a Hb of 5.6 and worsening symptoms before the patient became unresponsive.
What happened: Watery, blackish diarrhoea 4–5 times/day for 3 days, self-resolved. Family attributed colour to pomegranate juice.
What the advocate understood: Probably the juice. It stopped. No action needed.
What the advocate needed to know:
The missed red flag: Melaena in a patient with known H. pylori ulcers, severe anaemia, and a Hb that subsequently did not rise appropriately with transfusion is a direct clinical connection. The family had the information (black stool) but not the interpretive framework to know it was dangerous.
What happened: The family obtained blood transfusions on GP advice. Hb rose from 5.6 to 5.7 after 2 units. One more unit was given the next day.
What the advocate understood: She needs blood. We are giving blood. More blood will help.
What the advocate needed to know:
The question the advocate needed to ask: “Doctor, she had 2 units of blood and her count only went up by 0.1. Why is that? Where is the blood going?”
This question was never raised in the documented narrative. Had it been, it would have prompted the clinician to investigate active haemorrhage or haemolysis — a different and more urgent investigation pathway.
What happened: 10 days after the last transfusion, the patient became severely breathless and lost consciousness. The family then sought emergency care.
What the advocate understood: She got much worse suddenly. We had to go.
What the advocate needed: This inflection point should have been pre-empted by a defined escalation threshold given at Inflection Point 4. If the family had been told: “If she loses consciousness, or becomes suddenly much more breathless, or if she runs a temperature above 38°C or below 36°C — go to the emergency department without waiting” — the response time would have been faster.
The critical question is not whether the family acted appropriately in this moment (they did), but whether the 10-day interval between declining counts and loss of consciousness contained observable warning signs that the family saw but did not know were urgent. Based on the narrative, the answer is yes: worsening breathlessness, nausea, vertigo, and back pain had been present for at least the 10 days prior.
This section is not only an analysis of gaps. The family:
These competencies, in the absence of any structured training, are significant. They demonstrate that with appropriate information and a defined escalation framework, this family was capable of managing a complex care situation. The gaps were not in the family’s capacity — they were in the information they were given.
What this family would have needed — structured as the three questions from Module 2 Step 2.8:
[Remember] What is the main thing wrong with my family member?
“She has a long-standing spinal condition that is now getting worse, AND something is also affecting her blood and her organs. The doctors are trying to find out what is causing both problems.”
[Understand] Why is the current treatment important?
“The blood transfusions are replacing blood she is losing or that her body is destroying. The antibiotics are treating a blood infection. Neither of these treats the root cause — which is still being investigated. She will continue to need both until we find and treat the underlying problem.”
[Apply] If she develops any of these signs, what will you do immediately?
“If she loses consciousness, has a temperature below 36°C or above 38°C, has blood or black stool, or cannot breathe without the machine — I will call emergency services or go to the hospital immediately, and I will tell the doctor: ‘Her blood is not rising with transfusion, she has a stony hard spleen, and she has protein in her spinal fluid.’”
For the clinical team at first contact with pancytopenia: Before discharging with “wait until counts improve,” provide the family with: a written red-flag list, a specific return threshold (e.g. “if breathless, feverish, or unconscious — go immediately”), and a one-sentence explanation of why the counts are low if known.
For the prescribing clinician: When Baclofen frequency escalates, review renal and hepatic function before continuing. Educate the family that more Baclofen is not always safer — and define the signs of toxicity in terms they can observe at home.
For the advocate at discharge with an NGT: Provide the Module 2 Step 2.7 handover brief in written form. A 150-word plain-language summary of who the patient is, what her conditions are, what medications she is on, and what to do if she deteriorates — given to the family in writing — is the minimum information transfer for a high-risk home discharge.
For the healthcare system: When a patient with a multi-year diagnosis presents with new features that do not fit the existing diagnosis (pancytopenia, stony spleen), there should be a formal mechanism to prompt: “Is this the same condition or a new one?” This pivot was never triggered in this case despite four separate hospital encounters in the final four months of the patient’s life.
Document generated using the VibeRounds framework. All clinical content is educational. This is not a clinical guideline or diagnostic service. Independent clinical verification is required before acting on any content in this document.