The Unthinkable Question: Could Our Cure Cause More Harm?

Introduction

Every clinician who has ever started empirical low-molecular-weight heparin (LMWH) for a patient breathless in the Emergency Department knows the knot in the stomach: what if I am wrong, and the drug hurts more than it helps?
Pulmonary embolism (PE) remains one of the few acute medical emergencies in which we routinely administer potentially lethal anticoagulation before we prove the diagnosis. The therapeutic window is narrow; under-dose and we risk recurrent thrombosis and death, overdose and we court major bleeding, heparin-induced thrombocytopenia, and—overdosing becomes more dramatic in cases of CKD (Creatinine clearance <30mls/min) when renal-adjustment is ignored.

In 2010 the UK National Patient Safety Agency (NPSA) fired a warning shot: dosing errors with LMWH were causing “severe harm or death”. The signal was real, yet the national reporting systems captured only the tip of the iceberg. As the respiratory lead at Whiston Hospital, part of St Helens & Knowsley Teaching NHS Trust, I resolved to quantify the submerged bulk of that iceberg in my own institution. The results of our retrospective audit—70 consecutive patients investigated for suspected PE during September 2010—are sobering, and they underpin the reflections in this blog.

Why renal function matters more than we admit

LMWHs are hydrophilic, 4–6 kDa molecules predominantly cleared by the kidneys via glomerular filtration. Once creatinine clearance (CrCl) drops below 30 ml/min, accumulation of anti-Xa activity is near-linear, doubling the bleeding risk with each 12-hourly dose if standard therapeutic doses are continued. European and UK guidelines therefore recommend either:

• Dose reduction (enoxaparin 1 mg/kg once daily instead of 1 mg/kg twice daily), or
• Anti-Xa monitoring with target 0.6–1.0 IU/ml 4 h post-dose, or
• Switch to unfractionated heparin infusion.

Yet in our 2010 cohort, three patients had documented CrCl < 30 ml/min; only one had the LMWH dose adjusted. Two continued on full-dose enoxaparin for ≥48 h until pharmacy review—an oversight that, in the era of automated eGFR reporting, should never occur.

The audit in numbers

The apparent absence of harm is, I would argue, an artefact of poor ascertainment rather than proof of safety. None of the case notes contained a standardized bleeding score; renal function was not re-checked within 72 h of starting treatment; and no patient had anti-Xa levels measured. In short, we did not look systematically, so we did not find.

From signal to action: building a digital safety net in India

Paper protocols laminated to the drug chart have limited traction at 3 a.m. We therefore designed a smartphone/web app—“VTE-Safe”—that:

1. Integrates with the hospital’s electronic prescribing platform.
2. Auto-calculates CrCl using the validated CKD-EPI 2021 equation and flags CrCl < 30 ml/min in real time.
3. Suggests evidence-based dose adjustments (enoxaparin, dalteparin, tinzaparin) or switches to UFH.
4. Records every override together with the prescriber’s free-text justification—data that are encrypted and forwarded to the national database.
5. Follows the patient for 30 days, pushing reminders for platelet count (HIT surveillance) and renal function.

By deployment of this app, you would aim to reduce inappropriate prescription of low molecular weight heparin and create a monitoring tool for potential underlying side effects.

Practical checklist before you click “prescribe”

“Renal function, weight, clotting screen and platelets—every time, every patient, every dose.”

Before writing the first LMWH prescription for suspected PE, ask yourself:

• Have I documented a validated clinical probability score (Wells or revised Geneva)?
• Has the patient been weighed today on a calibrated scale, or am I guessing?
• What is the CrCl, and have I adjusted for BSA-indexation if the patient is cachectic or morbidly obese?
• Are there additive bleeding risks—recent surgery, cerebral metastases, platelets < 150 × 10⁹/l, concurrent antiplatelet therapy?
• Is anti-Xa monitoring required, and can the nursing staff comply with the 4-hour post-dose sampling window?
• Have I set a review date (≤ 72 h) to re-examine benefit–risk once imaging is reported?

If any answer is “no”, pause. A 6-hour delay while you gather data is preferable to a preventable intracranial haemorrhage.

Broader lessons for low- and middle-income settings

India, my country of origin, registers < 2 % of estimated adverse drug reactions (ADR). Under-reporting is multifactorial: medico-legal anxiety, lack of pharmacovigilance training, and no electronic trigger tools. The VTE-Safe model is open-source; the API can be grafted onto the Electronic Vaccine Intelligence Network (EVIN) backbone already used by the Universal Immunisation Programme. Capturing ADRs for high-risk drugs such as LMWH would cost < ₹3 per patient episode—cheaper than managing one retroperitoneal bleed in a private ICU.

Conclusion

Low-molecular-weight heparin remains the cornerstone of empirical therapy for suspected PE, but its narrow therapeutic index demands respect. Our 2010 audit revealed that more than half of patients received incorrect dosing, and system blindness gave false reassurance of safety. Digital decision support, mandatory renal-function checks and a culture that celebrates not giving a drug when it is unsafe are the only sustainable ways to turn that statistic around. Until national reporting systems achieve near-real-time granularity, the onus is on individual prescribers to ask one simple question every time they sign the chart: if the CTPA comes back negative tomorrow, will what I prescribed today still be defensible?