Thrombocytopenia: nine teaching cases.

Topic Pseudothrombocytopenia·Curriculum Platelet disorders — laboratory recognition of thrombocytopenia; artefact; peripheral blood film interpretation; safe exclusion of spurious results.

Model answer

Diagnosis. EDTA-dependent pseudothrombocytopenia. The patient has no bleeding phenotype, the analyser and film show platelet aggregation, and the citrate count is normal.

Laboratory lesson. Always confirm an unexpected isolated thrombocytopenia before labelling a patient with ITP or another disease. Examine the blood film, review the analyser flags and, when appropriate, obtain a repeat sample in citrate or another validated anticoagulant. A citrate result may require correction for dilution according to the local laboratory method.

Clinical consequence. The key hazard is not bleeding from pseudothrombocytopenia but iatrogenic harm from an incorrect diagnosis: unnecessary steroids, IVIg, bone marrow examination, surgical delay or platelet transfusion. The patient should have a clear laboratory comment so that the finding is not repeatedly rediscovered as a "new thrombocytopenia".

Key learning points

• An unexpected isolated low platelet count should prompt confirmation before treatment.
• A blood film is a critical part of thrombocytopenia assessment, not an optional extra.
• EDTA-dependent platelet aggregation can cause substantial analyser under-counting.
• Citrate or another validated alternative specimen can establish the true platelet count.
• Document the artefact clearly to prevent repeated investigations.

Further reading

• Cattaneo M. Pseudothrombocytopenia and spuriously low platelet counts. Haematologica 2025 (open access)
• BSH — Clinical and laboratory diagnosis of heritable platelet disorders (2021)

Topic Primary immune thrombocytopenia·Curriculum Platelet disorders — diagnosis of ITP; bleeding risk assessment; first-line corticosteroids/IVIg; treatment thresholds and follow-up.

Model answer

Diagnosis. Newly diagnosed primary ITP is the most likely diagnosis after excluding secondary causes and important mimics. In a patient with isolated thrombocytopenia, the examination and blood film are as important as the numerical count.

Immediate management. With platelets 7 × 10⁹/L and mucosal bleeding, treatment is indicated. A corticosteroid regimen is appropriate; IVIg can be added when a rapid response is needed. Investigate secondary causes according to the clinical setting, including HIV and HCV and selected autoimmune, immunodeficiency or lymphoproliferative disorders.

The febrile mimics come first in this setting. In India, acute febrile thrombocytopenia should prompt active consideration of dengue, malaria, enteric fever and other infections before immunosuppression is commenced. What makes ITP the working diagnosis here is the complete absence of fever throughout the illness, together with a normal haemoglobin, a normal white count and a film showing large platelets and nothing else. Had she been febrile, an NS1 antigen or dengue serology would precede any thought of corticosteroids — giving steroids to a patient with an undiagnosed acute febrile illness is a real and avoidable harm.

Treatment philosophy. The objective is a safe platelet count, not a normal platelet count. Avoid repeated short courses of corticosteroid when a patient becomes steroid-dependent; second-line options include a TPO receptor agonist, rituximab or splenectomy, chosen according to disease duration, bleeding history, comorbidity, patient preference and local availability.

A note on access, and on guidance. Corticosteroids and IVIg are the practical first-line options in most Indian centres. The cost of IVIg often decides whether it is used, and rituximab and thrombopoietin receptor agonists are variably available and frequently self-funded. Availability and cost usually shape the second-line treatment strategy, and it is worth an early discussion with the patient and family rather than one at the point of failure. ASH 2026 guidance conditionally favours combination initial therapy with corticosteroids plus either a TPO-receptor agonist or rituximab in adults with primary ITP, instead of corticosteroids alone. IVIg is appropriate when a rapid platelet increment is required. The ASH 2019 guideline remains the current ASH source for children, for diagnosis and for emergency management.

Key learning points

• ITP is a diagnosis of exclusion supported by isolated thrombocytopenia and an appropriate blood film.
• Treat bleeding risk rather than pursuing a normal platelet count.
• Corticosteroids are first-line; IVIg is useful when a rapid platelet increment is required.
• Platelet transfusion is reserved for serious/life-threatening bleeding or urgent haemostatic situations and should be combined with ITP-directed treatment.
• Second-line therapy should be individualised; splenectomy is no longer the default early option.
• In India, exclude dengue and other acute febrile illnesses before treating severe thrombocytopenia as ITP — steroids given blind to an undiagnosed febrile illness cause harm.
• Second-line choice in practice is often decided by cost and availability, not by algorithm. Raise it with the family early.

Further reading

• Provan D et al. Updated international consensus report on primary ITP. Blood Advances 2019 (open access)
• ASH 2026 ITP guidelines — initial and second-line therapy in adults
• ASH 2019 ITP guidelines — still current for children, diagnosis and emergency management. Blood Advances 2019

Topic Thrombotic microangiopathy·Curriculum Platelet disorders — thrombotic microangiopathy; sequencing investigations in a haematological emergency; PLASMIC score; ADAMTS13; congenital and immune TTP; plasma exchange, caplacizumab and immunosuppression.

Model answer

How the case should unfold. A woman with confusion, thrombocytopenia and anaemia is a thrombotic microangiopathy until the film says otherwise, and the film is the investigation that earns its place first. Everything that follows — the haemolysis screen, the coagulation screen, the DAT, ADAMTS13 — refines a diagnosis you should already be treating.

Establishing the microangiopathy. Schistocytes with a raised LDH and unconjugated bilirubin, undetectable haptoglobin and a reticulocytosis establish mechanical haemolysis. The negative DAT argues strongly against immune destruction, though a small minority of AIHA is DAT-negative. Normal PT, APTT and fibrinogen argue firmly against DIC — the single most useful discriminator at the bedside and the one most often skipped — without excluding early or compensated DIC. At 14 weeks, HELLP does not fit.

Using PLASMIC properly. The score exists to support a decision you must make before the assay returns. Seven components, one point each: platelets below 30 × 10⁹/L, haemolysis, no active cancer, no transplant, MCV below 90 fL, INR below 1.5, creatinine below 2.0 mg/dL. Scores of 6–7 carry a high probability of severe ADAMTS13 deficiency. Use it to justify starting, not to justify waiting — and never let a low score alone stop you treating a patient who looks like TTP.

Treatment of the acute episode. Urgent plasma exchange plus corticosteroids, started on clinical grounds. Caplacizumab is added in the acute immune episode where available: it blocks the VWF A1 domain, shortens time to platelet recovery and reduces exacerbations, at the cost of bleeding risk, and it is stopped before invasive procedures. Note that pregnancy data are limited to case reports and small series, so its use in pregnancy is off-label and generally reserved for refractory disease. Rituximab suppresses the autoantibody and reduces relapse. Avoid platelet transfusion unless there is life-threatening bleeding — the microthrombi are platelet-rich, and transfusing fuels them.

Congenital TTP, and why it belongs in this case. Severe ADAMTS13 deficiency with no detectable antibody or inhibitor should prompt gene sequencing. Congenital TTP often declares itself for the first time in pregnancy, because von Willebrand factor climbs through gestation and exhausts the residual enzyme activity; the history that gives it away is neonatal jaundice needing exchange transfusion, unexplained childhood thrombocytopenia, or a previous pregnancy loss treated with plasma. Ask about all three in any young woman presenting with a microangiopathy. Management then diverges sharply: the enzyme has to be replaced, and immunosuppression has no part in it. Caplacizumab is not licensed for congenital TTP. For maintenance in remission the 2025 ISTH focused update makes a strong recommendation for recombinant ADAMTS13 over prophylactic plasma infusion. In pregnancy, where experience with the recombinant product is still limited, regular prophylactic plasma infusion remains the recommended approach — and she needs it for the remainder of this pregnancy and in any future one. One caution: a negative antibody is much weaker evidence than it looks — up to a quarter of immune cases have no detectable anti-ADAMTS13 IgG at presentation, and a negative functional inhibitor is commoner still. Genetics, not the antibody result, makes the call.

The obstetric dimension. Untreated, TTP in pregnancy carries a high risk of fetal loss through placental infarction, which is the likely explanation for her previous stillbirth. Manage jointly with obstetrics and with a centre that has experience of TTP in pregnancy, and plan prophylaxis for the remainder of this pregnancy from the point the diagnosis is made.

Key learning points

• In suspected TTP the blood film is the first investigation, not ADAMTS13 — it is immediate and it answers the question that decides treatment.
• Normal PT, APTT and fibrinogen are what separate a microangiopathy from DIC. Check them early.
• PLASMIC (0–4 low, 5 intermediate, 6–7 high) supports treating before ADAMTS13 returns. It never replaces the assay, and a low score does not overrule a convincing clinical picture.
• Plasma exchange plus corticosteroids is the backbone; caplacizumab shortens platelet recovery in immune TTP but carries a bleeding risk, and it is not licensed for congenital disease, which is treated by replacing the enzyme.
• Congenital TTP is bimodal — about half presents in the first years of life and half in adulthood, and in women pregnancy is the commonest adult trigger. Roughly a quarter of pregnancy-onset TTP with severe ADAMTS13 deficiency proves to be congenital.
• Ask every young woman with a microangiopathy about neonatal jaundice, childhood thrombocytopenia and previous pregnancy loss treated with plasma.
• Avoid platelet transfusion unless bleeding is life-threatening.

Further reading

• ISTH — Guidelines for the diagnosis of TTP. J Thromb Haemost 2020
• ISTH — Guidelines for the treatment of TTP. J Thromb Haemost 2020
• ISTH — 2025 focused update of the TTP guidelines. J Thromb Haemost 2025;23:3711–32
• Congenital TTP (Upshaw–Schulman) — presentation and management in pregnancy, covered in the ISTH 2025 focused update

Topic Heparin-induced thrombocytopenia·Curriculum Platelet disorders — heparin-induced thrombocytopenia; applying the 4Ts score before laboratory testing; PF4 immunoassay and functional assays; non-heparin anticoagulation and the timing of warfarin.

Model answer

The question behind the phone call. “Should I send a HIT screen?” is the wrong first question. The right one is what the pre-test probability is, because that determines both whether the test means anything and whether heparin stops today.

Diagnosis. Probable HIT with thrombosis (HITT). The key clues are the timing, magnitude of platelet fall and new thrombosis following heparin exposure. Cardiac surgery is a setting in which early postoperative thrombocytopenia is common, so the second platelet fall is particularly important.

Immediate action. Stop unfractionated and low-molecular-weight heparin and commence a non-heparin anticoagulant. Options include argatroban, bivalirudin, fondaparinux or a DOAC in appropriate patients, guided by renal/hepatic function, bleeding risk, procedural plans and local protocol. In HIT without clinically apparent thrombosis, ultrasound of the lower limbs should be considered because occult DVT is common and changes treatment duration.

Laboratory interpretation. The PF4/heparin immunoassay is sensitive but not perfectly specific. A strong positive result in a high-4Ts patient supports the diagnosis; functional platelet activation testing can improve specificity. A low 4Ts score should not trigger indiscriminate testing or empiric non-heparin anticoagulation.

Choosing the non-heparin anticoagulant. Not all the alternatives are equal here. Seven days after cardiac surgery he is at high bleeding risk, and argatroban or bivalirudin are preferred in that situation precisely because their effect is short-lived and can be stopped quickly. Danaparoid and fondaparinux are alternatives but have long half-lives and are not readily reversible; avoid them in severe renal impairment (CrCl below 30 mL/min) and use them with caution and anti-Xa monitoring in moderate impairment. Direct oral anticoagulants are reasonable in the clinically stable patient at average bleeding risk, but experience in limb- or life-threatening thrombosis is limited. Warfarin when indicated should not be started until platelet recovery — conventionally to 150 × 10⁹/L or above — and overlap for at least five days. Warfarin initiation during active HIT without platelet count recovery risks venous limb gangrene and skin necrosis.

Key learning points

• Score the 4Ts before sending the immunoassay. The PF4 assay is sensitive but poorly specific, especially after cardiac surgery, and a result read without a pre-test probability misleads.
• 4Ts domains score 0–2 each, maximum 8: fall in platelets, timing, thrombosis, other causes. 6–8 is high probability, 4–5 intermediate, 0–3 low.
• In a post-cardiac-surgery patient the “other causes” domain is the soft one — bypass itself causes an early fall, but a fresh fall on day 7 is not explained by the operation.
• Stop every source of heparin, including line flushes and heparin-bonded catheters.
• Stopping heparin is not enough: HIT is prothrombotic, so start a therapeutic-dose non-heparin anticoagulant.
• Argatroban or bivalirudin where bleeding risk is high. Danaparoid and fondaparinux have long half-lives, are not readily reversible, and should be avoided in severe renal impairment (CrCl below 30 mL/min) and used cautiously with anti-Xa monitoring in moderate impairment.
• Do not initiate warfarin during acute HIT; start it after platelet recovery, usually ≥150 × 10⁹/L, with adequate overlap — premature warfarin can precipitate venous limb gangrene and skin necrosis.
• Consider lower-limb ultrasound even without symptoms: occult DVT is common and changes the duration of anticoagulation.
• Avoid prophylactic platelet transfusion; reserve it for serious bleeding.

Further reading

• ASH 2018 guidelines for VTE — heparin-induced thrombocytopenia
• BSH — Diagnosis and management of heparin-induced thrombocytopenia, third edition (2023, reviewed 2024)

Topic Gestational thrombocytopenia / ITP / pre-eclampsia·Curriculum Platelet disorders in pregnancy — distinguishing gestational thrombocytopenia, ITP and hypertensive disorders; delivery planning; neuraxial anaesthesia; neonatal assessment.

Model answer

Diagnosis. Gestational thrombocytopenia is most likely. It is typically mild, appears in later pregnancy, is not associated with maternal bleeding or other cytopenias, and resolves postpartum. Platelets below about 70 × 10⁹/L are less typical and should trigger a careful search for alternative diagnoses, particularly ITP or a hypertensive/TMA syndrome.

Delivery planning. Management depends on the maternal diagnosis, trend, bleeding history and planned mode of delivery. In known/suspected ITP, NICE recommends an obstetric-led setting with neonatal support and neonatal precautions because maternal platelet count does not reliably predict neonatal platelet count.

Neuraxial anaesthesia. Use the contemporary NICE thresholds as a guide rather than treating one numerical count as an absolute universal rule. Above 80 × 10⁹/L is generally acceptable to consider; 50–80 × 10⁹/L requires multidisciplinary, individualised assessment; below 50 × 10⁹/L neuraxial anaesthesia is generally avoided.

Key learning points

• Gestational thrombocytopenia is a diagnosis of exclusion and is usually mild and late in pregnancy.
• Platelets below about 70 × 10⁹/L should prompt careful reassessment of the diagnosis.
• New hypertension, haemolysis or liver enzyme abnormalities require urgent consideration of pre-eclampsia/HELLP or TMA.
• Maternal platelet count does not predict neonatal platelet count reliably in ITP.
• Neuraxial anaesthesia decisions should follow current obstetric/anaesthetic guidance and multidisciplinary assessment.

Further reading

• NICE NG121 — Intrapartum care for women with existing medical conditions
• Provan D et al. International consensus report on primary ITP (pregnancy section). Blood Advances 2019
• Fogerty AE. ITP in pregnancy: diagnostics and therapeutics. ASH Education Program 2024 (free)

Topic Drug-induced immune thrombocytopenia·Curriculum Platelet disorders — drug-induced thrombocytopenia; diagnosis by temporal relationship; drug cessation; recurrence risk; platelet transfusion and IVIg in severe bleeding.

Model answer

Diagnosis. Probable vancomycin-induced immune thrombocytopenia. The platelet nadir is profound, the fall is isolated, there is a strong temporal association and a history suggesting re-exposure. The normal coagulation profile and absence of schistocytes argue against DIC and TMA.

Management. Stop vancomycin immediately and select an alternative antimicrobial without the same risk where possible. Document the adverse drug reaction prominently; future re-exposure may provoke rapid, life-threatening thrombocytopenia. Drug-dependent platelet antibody testing can be helpful retrospectively in specialist laboratories but should not delay withdrawal of the drug.

Bleeding. With active mucosal bleeding and a platelet count of 4 × 10⁹/L, treat urgently. IVIg can produce a relatively rapid increase in platelet count. Corticosteroids are often given, though benefit is uncertain and warrants caution here given active prosthetic joint infection. The pathophysiology here is a drug-dependent antibody and recovery follows drug clearance. The most important intervention is withdrawal of the culprit medication. Note also that oral blood blisters with a count of 4 × 10⁹/L are wet purpura, which is serious bleeding and warrants urgent haemostatic management including platelet transfusion. Replace vancomycin with an effective non-cross-reacting antimicrobial in consultation with microbiology/infectious diseases, but avoid linezolid, as it can cause myelosuppression and thrombocytopenia itself. Platelet transfusion can be life-saving in severe bleeding even when the increment is transient.

The drugs to know. Several hundred medicines have been implicated, but a short list accounts for most well-documented cases and is worth carrying in your head. Antimicrobials: vancomycin, co-trimoxazole and other sulphonamides, rifampicin, linezolid, piperacillin, ceftriaxone. Anticonvulsants: carbamazepine, phenytoin, sodium valproate. Quinine and cardiovascular drugs: quinine and quinidine — quinine remains among the most frequently implicated of all, and patients rarely volunteer tonic water or a traditional remedy unless you ask — plus the glycoprotein IIb/IIIa inhibitors abciximab, eptifibatide and tirofiban, which cause thrombocytopenia within hours. Others: NSAIDs including ibuprofen and naproxen, oxaliplatin, penicillamine, thiazide diuretics, and heparin, which is immune-mediated but by a wholly different mechanism. The University of Oklahoma database linked below is the standard reference when a less familiar drug is in the frame.

Key learning points

• Drug-induced immune thrombocytopenia can cause platelet counts in single digits.
• The temporal relationship and previous exposure are often more useful than any single laboratory test.
• Stop the suspected drug immediately; do not wait for specialised antibody testing.
• Document the culprit drug clearly because re-exposure can be dangerous.
• In major bleeding, platelet transfusion may be used despite immune destruction, alongside other haemostatic therapy.
• Know the usual suspects: vancomycin, co-trimoxazole, rifampicin, linezolid, carbamazepine, phenytoin, valproate, quinine and quinidine, GPIIb/IIIa inhibitors, NSAIDs, oxaliplatin. Ask specifically about quinine in tonic water and in traditional remedies.

Further reading

• Aster RH, Bougie DW. Drug-induced immune thrombocytopenia. N Engl J Med 2007 (subscription)
• Platelets on the Web (University of Oklahoma) — drug-induced thrombocytopenia database, reports to 2018

Topic Hypersplenism and cirrhosis·Curriculum Platelet disorders — thrombocytopenia in chronic liver disease; hypersplenism; thrombopoietin deficiency; procedure-related platelet targets.

Model answer

Diagnosis. Chronic thrombocytopenia related to advanced liver disease with portal hypertension and splenomegaly, likely contributed to by both hypersplenism and reduced thrombopoietin production. The associated leukopenia supports a broader sequestration/marrow effect rather than isolated ITP.

Haemostatic interpretation. Cirrhosis does not simply equal an anticoagulated state. Both procoagulant and anticoagulant proteins are reduced, platelet function changes, and endothelial factors are altered. INR should not be used as a stand-alone bleeding-risk test or as a universal trigger for plasma.

Procedural management. There is no single platelet threshold suitable for every procedure. The decision depends on bleeding risk of the procedure, platelet trend, fibrinogen, active bleeding, renal dysfunction, infection and whether a reliable local pathway exists. Where a planned high-risk procedure requires platelet augmentation, TPO receptor agonists or platelet transfusion may be considered depending on timing and setting.

The request you have been sent is the teaching point. A pre-emptive order for platelets and plasma to 'cover' a procedure in cirrhosis is common, and usually wrong. Plasma does not reliably correct the INR in liver disease, volume-loads a patient with portal hypertension and can raise portal pressure, and the INR was never validated as a bleeding predictor in this population. Where the platelet count genuinely needs raising for a higher-risk procedure, a thrombopoietin receptor agonist given in advance is an alternative to transfusion where it is available and affordable. The most useful thing you can do is discuss the specific procedure and its real bleeding risk with the proceduralist rather than issue products against a number.

Give the proceduralist a number and an alternative route. A platelet count of 62 × 10⁹/L needs no correction for a diagnostic liver biopsy under any current threshold — the usual procedural cut-off sits nearer 50 × 10⁹/L, and that is the concrete answer to give rather than a general argument about rebalanced haemostasis. Transjugular biopsy is often preferred in significant ascites and/or coagulopathy in view of bleeding risk and also allows hepatic venous pressure gradient measurement.

Key learning points

• Thrombocytopenia in cirrhosis is multifactorial; hypersplenism is not the whole explanation.
• A prolonged INR does not mean that a patient with cirrhosis is simply "auto-anticoagulated".
• Bleeding and thrombosis can both occur in advanced liver disease.
• Do not use a universal platelet threshold for every procedure.
• Use procedure-specific hepatology/transfusion guidance and avoid unnecessary correction of laboratory abnormalities.
• A routine request for platelets and plasma before a procedure in cirrhosis usually reflects the number, not the risk. Discuss the specific procedure rather than transfusing to a target.

Further reading

• EASL — Prevention and management of bleeding and thrombosis in patients with cirrhosis (2022)
• AASLD — Vascular liver disorders, portal vein thrombosis and procedural bleeding: practice guidance (2020)

Topic Acute promyelocytic leukaemia and DIC·Curriculum Platelet disorders — thrombocytopenia in acute leukaemia; DIC; acute promyelocytic leukaemia; urgent ATRA; transfusion support.

Model answer

Diagnosis. Suspected acute promyelocytic leukaemia complicated by disseminated intravascular coagulation/marked consumptive coagulopathy. The combination of abnormal promyelocytes, Auer rods, profound thrombocytopenia, low fibrinogen and high D-dimer is highly characteristic.

Immediate treatment. Start ATRA immediately when APL is suspected; do not wait for molecular confirmation. Arrange urgent rapid PML::RARA testing and manage in close liaison with a specialist acute leukaemia service. APL-associated coagulopathy can deteriorate rapidly, particularly around invasive procedures.

Supportive care. Maintain vigorous blood product support, commonly aiming for platelets around 30–50 × 10⁹/L or higher in active bleeding and fibrinogen >1.5 g/L, with local protocol determining exact targets. Avoid unnecessary invasive procedures, especially lumbar puncture, while the coagulopathy is uncontrolled. Platelet transfusion is a core part of management in this setting, unlike in uncomplicated TTP.

What to do before transfer — the question that decides the outcome. Start ATRA now, before she moves and long before PML::RARA is confirmed. Early death from haemorrhage, most often intracranial or pulmonary, is the dominant cause of mortality in APL and it happens in the first hours to days, which in practice means during exactly the period this patient is being referred and transported. ATRA carries little cost if the diagnosis later proves wrong; withholding it while awaiting confirmation is the error that kills. Correct the coagulopathy aggressively with cryoprecipitate or fibrinogen concentrate and platelets before and during transfer, avoid all non-essential invasive procedures including central lines and lumbar puncture, and speak directly to the receiving haematologist rather than sending her with a letter. Because she has a headache with platelets of 14 × 10⁹/L and a fibrinogen of 0.9 g/L, urgent CT brain should be considered if immediately available, but this must not delay ATRA, correction of coagulopathy or transfer. Intracranial haemorrhage is the leading early killer in APL and a four-hour road transfer is exactly when it declares itself.

Key learning points

• Suspected APL is an emergency: start ATRA immediately.
• The major early threat is haemorrhage and/or thrombosis from APL-associated coagulopathy.
• Support platelets and fibrinogen proactively while definitive diagnosis and treatment are arranged.
• Morphology can be highly suggestive, but PML::RARA testing establishes the diagnosis.
• Do not apply the "avoid platelet transfusion" rule from TTP to APL-associated DIC.
• Start ATRA on clinical suspicion, before molecular confirmation and before transfer. Most APL deaths are early haemorrhagic deaths, and they occur during precisely the delay a referral creates.
• Avoid non-essential invasive procedures — central lines, lumbar puncture — until the coagulopathy is corrected.

Further reading

• Sanz MA et al. Management of acute promyelocytic leukaemia: updated ELN recommendations. Blood 2019
• BSH — Guidelines for the use of platelet transfusions (2017, reviewed 2022)

Topic Dengue·Curriculum Platelet disorders — thrombocytopenia in acute febrile illness; recognition of the critical phase of dengue; evidence-based use of platelet transfusion.

Model answer

Diagnosis. Dengue with warning signs, entering the critical phase. NS1 is positive, and the haematocrit has risen from 43% to 53% while the platelet count has fallen to 18 × 10⁹/L — the combination of rising haematocrit and falling platelets is itself a warning sign. The vomiting, abdominal pain, tachycardia and ultrasound evidence of plasma leakage complete the picture — his pulse pressure is still 38 mmHg, so he is not yet in compensated shock.

Read the timeline, not the platelet count. Dengue runs in three phases: febrile (2–7 days), critical (beginning around defervescence, typically days 3–7 of illness and lasting 24–48 hours) and recovery (a further 2–3 days). Deterioration clusters at the start of the critical phase, which is exactly when the patient and often the referring doctor believe things are improving because the fever has broken. Recognising that transition is the core clinical skill in this case.

What actually threatens him. Plasma leakage and hypovolaemic shock, not haemorrhage. A haematocrit rising by more than about 20% above baseline indicates significant leak. Watch the pulse pressure. In dengue it narrows before the systolic pressure falls, and a pulse pressure of 20 mmHg or less with tachycardia is compensated shock — severe dengue — requiring an immediate isotonic crystalloid bolus rather than maintenance fluid. His is 38 mmHg at present. By the time the systolic pressure drops, compensation has already failed.

Management. Admit. Titrated intravenous crystalloid guided by serial haematocrit, pulse, blood pressure, pulse pressure and hourly urine output, aiming for adequate perfusion rather than a normal haematocrit. Reduce fluids as the critical phase ends — fluid overload during reabsorption is a common iatrogenic cause of deterioration. Monitor the platelet count for trend, not as a transfusion trigger. Avoid NSAIDs and intramuscular injections; paracetamol for fever.

Transfusion. No prophylactic platelets. Transfuse for clinically significant bleeding, and consider it at a very low count in the narrow circumstances Indian national guidance sets out. Explain to the family, plainly and without dismissing their worry, that the platelet count is a marker of where he is in the illness rather than the thing that will harm him, and that the transfusion they are asking for carries risk without demonstrated benefit. This conversation is difficult and extremely common; having it well is part of the job.

Discharge. Afebrile for 48 hours without antipyretics, improving appetite and clinical state, stable haematocrit, adequate urine output, no respiratory distress and a rising platelet count. The platelet count typically recovers within 3–5 days of defervescence and does not need to be normal before discharge.

Key learning points

• Defervescence marks the start of the critical phase, not recovery. Deterioration happens as the fever settles.
• Rising haematocrit with falling platelets is a warning sign, and haematocrit trend is the key measurement — record a baseline early.
• A narrowing pulse pressure precedes hypotension in dengue shock; by the time systolic pressure falls, compensation has failed.
• Prophylactic platelet transfusion did not reduce bleeding and caused more adverse events in a randomised trial (Lye et al., Lancet 2017); NCVBDC 2023 and WHO 2025 both advise against it in the absence of bleeding.
• Titrated fluid is the treatment; reduce it during the reabsorption phase to avoid pulmonary oedema.
• Avoid NSAIDs and intramuscular injections. Paracetamol only.

Further reading

• WHO — Guidelines for clinical management of arboviral diseases: dengue, chikungunya, Zika and yellow fever (2025)
• NCVBDC / MoHFW India — National Guidelines for Clinical Management of Dengue Fever (2023)
• Lye DC et al. Prophylactic platelet transfusion in adult dengue: randomised trial. Lancet 2017