Sometimes the best antibiotic prescription is the one you don’t write.

Most unnecessary antibiotics don’t begin with negligence. They begin with good medicine—or at least, the sincere intention of practicing vigilant, responsive care. A rising white cell count. A positive culture report. A patient who appears “just unwell enough” to trigger clinical unease. That discomfort, born from years of training to err on the side of caution, is precisely where over-treatment often takes root.

In a single week, three patients drove home a principle that is straightforward in theory but profoundly challenging in daily practice: The correct antibiotic is sometimes no antibiotic at all.

Case 1: DKA, Leukocytosis, and the Illusion of Infection

A young woman in her 20s was admitted with severe diabetic ketoacidosis (DKA) after abruptly stopping her insulin. On arrival, she was profoundly acidotic (pH <7.0), confused, tachycardic, and hemodynamically unstable. She required immediate ICU admission, fluid resuscitation, insulin infusion, electrolyte repletion, and eventual intubation.

Standard DKA management unfolded. By day 5, she was improving metabolically—but new lab results appeared:

• WBC count climbed to 20 × 10⁹/L
• CRP was rising
• Urine culture (via indwelling Foley) returned positive

The ward round narrative shifted instantly from metabolic correction to broad-spectrum antibiotics. The presence of bacteria seemed to eclipse the absence of a true infectious syndrome.

The Infectious Diseases consult recommended no antibiotics.
Reasoning: DKA itself causes marked leukocytosis and CRP elevation through stress hormones and cytokine release. The Foley catheter had predictably led to asymptomatic bacteriuria. The patient had no fever, no focal symptoms, and no clinical deterioration. Antibiotics were withheld. As the DKA resolved, the WBC and CRP normalized without any antimicrobial intervention.

Case 2: Hypotension, the Reflex of Ceftriaxone, and the Pulmonary Embolism

A few days later, an older woman developed sudden hypotension, tachycardia, and mild leukocytosis. Within minutes, ceftriaxone was started—following the well-trained reflex that links hypotension + tachycardia + leukocytosis with sepsis.

But reflexes occasionally misfire.

This patient had recently been diagnosed with pulmonary embolism. Her oxygen saturation was drifting downward. She had a history of recurrent thrombosis.

The Infectious Diseases consult recommended pausing the antibiotics and reconsidering the story. Repeat imaging confirmed another pulmonary embolism. Once anticoagulation was properly restarted and therapeutic levels achieved, her blood pressure stabilized and oxygenation improved.

The antibiotics had been treating bacteria that were never there.

Case 3: The Suprapubic Catheter and the Antimicrobial Tasting Menu

The third patient had a long-term suprapubic catheter. His urine cultures had developed a personality of their own—growing bacteria enthusiastically. Occasionally blood cultures joined the celebration.

Each positive result triggered another antibiotic:

• Levofloxacin
• Then piperacillin-tazobactam
• Then linezolid
• Then meropenem

At one point the drug chart looked less like a treatment plan and more like an antimicrobial tasting menu.

Yet the patient himself remained perfectly well. No fever. No hypotension. No local signs of infection. No rigors. No flank pain. Just a catheter quietly collecting microbes, as long-term catheters tend to do.

The Infectious Diseases consult suggested something radical: Stop everything. Observe.

Long-term urinary catheters are almost always colonized. Bacteriuria alone does not equal infection. Guidelines are unequivocal—do not treat asymptomatic bacteriuria in patients with chronic indwelling catheters.

The antibiotics were stopped. Nothing happened.
Which, in this case, was exactly what we wanted.

---

Three patients. Three moments where the reflex said treat, and the correct answer was stop.

Medicine trains us to intervene. To prescribe something that signals action. But every unnecessary antibiotic exerts pressure — on the patient’s microbiome, on the ward’s ecology, and eventually on the hospital’s resistance patterns.

A colleague once told me: “The hardest antibiotic to prescribe is the one you don’t write.”
Antibiotics treat infections. They do not treat anxiety.

Why We Overprescribe: Systemic and Psychological Drivers

This is rarely ignorance. It stems from:

• Fear of missing sepsis
• Overreliance on laboratory data
• The psychological power of a positive culture report
• Defensive medicine and time pressure

The Real Costs of “Just in Case” Therapy

Patient-level harms: allergic reactions, nephrotoxicity, C. difficile colitis, microbiome disruption.
System-level harms: accelerating antimicrobial resistance, prolonged hospital stays, increased costs.
Diagnostic harms: anchoring bias that masks the true diagnosis (DKA, pulmonary embolism, etc.).

A Practical Framework: Pause Before Prescribing

Before starting antibiotics, ask:

1. Is there a clinical syndrome of infection (fever, localizing symptoms, unexplained deterioration)?
2. Does the culture represent true infection or colonization?
3. Could this be sterile inflammation (DKA, PE, trauma, pancreatitis)?
4. What happens if I wait 6–12 hours with close observation in a stable patient?

Bottom Line

• Not every elevated WBC demands broad-spectrum antibiotics.
• Not every positive culture equals disease.
• Not every unstable patient has bacterial sepsis.

Sometimes, the highest level of care is not escalation—it is thoughtful, evidence-based restraint.

In the age of rising resistance, true clinical mastery lies not only in knowing when to start therapy, but in having the wisdom and courage to withhold it when it is not needed. That restraint saves lives too.