Urinary Tract Infections: Choosing and Timing the Antibiotic

First, classify — the diagnosis chooses the drug

"UTI" is not one disease. Four clinical pictures sit under the label, and each is managed differently. Uncomplicated cystitis is a lower-tract (bladder) infection in an otherwise healthy, non-pregnant adult with a normal urinary tract — burning, frequency, urgency, suprapubic discomfort. Complicated or upper-tract infection means the infection has reached the kidney (pyelonephritis: fever, flank pain, systemic illness), or the patient has a structural or functional problem (obstruction, stones, catheter, immunosuppression, poor kidney function) that makes failure and progression more likely. Asymptomatic bacteriuria is bacteria in the urine with no symptoms at all — and, crucially, in most people it should not be treated. Catheter-associated UTI is its own category, because a catheter is always colonised and the urine is never sterile. Naming which of the four you are looking at is the single most important step, because it decides whether to treat, which drug reaches the site, and how long to give it.

The organisms — know your enemy

Empirical therapy works because the culprits are predictable. Escherichia coli — a gut coliform that ascends the short female urethra — causes the large majority of community UTIs. Behind it come the other Enterobacterales: Klebsiella, Proteus (a urea-splitter that raises urinary pH and favours struvite stones), and Enterococcus. In young, sexually active women a second name matters: Staphylococcus saprophyticus, a coagulase-negative staphylococcus that is a classic cause of cystitis in that group. Because E. coli dominates so heavily, guidelines let you start an antibiotic on the clinical picture alone in simple cystitis, sending urine for culture only when the diagnosis is uncertain, the infection is complicated or upper-tract, the patient is pregnant, or the first drug fails. The predictability of the bug is exactly what makes short, targeted, culture-free treatment safe.

First-line drugs for cystitis — and why they suit the bladder

The two workhorses are chosen precisely because they concentrate in urine and spare the rest of the body. Nitrofurantoin is first-line for uncomplicated cystitis. The kidney concentrates it into the urine, where it is reduced by bacterial enzymes into reactive intermediates that damage bacterial DNA, ribosomes and metabolism at multiple points at once — a scattergun mechanism that has kept resistance strikingly low over decades. But those same properties define its limits. It achieves useful levels only in urine, never in tissue or blood, so it is useless in pyelonephritis and must never be used when infection has left the bladder. It is contraindicated in significant renal impairment (broadly an eGFR below about 45 mL/min): failing kidneys can't concentrate it into the urine, so it fails to work while accumulating systemic toxicity (peripheral neuropathy, and with long-term use pulmonary fibrosis and hepatotoxicity). And it is avoided at term and in the neonatal period because it can trigger haemolysis in the newborn — dangerous in G6PD deficiency, a link back to the Haematology chapter, where nitrofurantoin sits on the standard list of oxidant drugs that haemolyse G6PD-deficient red cells.

Trimethoprim is the main first-line alternative. It is a folate antagonist: it inhibits bacterial dihydrofolate reductase, starving the organism of the tetrahydrofolate it needs to make DNA — the same folate-pathway logic taught in the Antimicrobials section, where trimethoprim is often paired with a sulfonamide as co-trimoxazole to block two consecutive steps. Two cautions dominate its use. It is teratogenic in the first trimester: by blocking folate it raises the risk of neural tube defects, so it is avoided in early pregnancy (and folate supplementation matters if it must be used later). And because it structurally resembles amiloride, it blocks the distal renal sodium channel and reduces potassium excretion — so it can cause hyperkalaemia, especially in older patients, in renal impairment, or alongside ACE inhibitors, ARBs or potassium-sparing diuretics. It also raises creatinine by blocking its tubular secretion (often a benign, reversible rise) and potentiates warfarin. Rising E. coli resistance to trimethoprim in many regions is now steering first-line choice back toward nitrofurantoin.

Two more urine-directed options round out the cystitis toolkit. Fosfomycin, given as a single 3 g oral dose, inhibits an early step in bacterial cell-wall synthesis and reaches very high urinary concentrations; its single-dose convenience and activity against many resistant organisms — including some ESBL-producing E. coli — make it valuable when resistance limits the usual choices. Pivmecillinam, a penicillin (an amidinopenicillin) with particular Gram-negative activity, is another well-established first-line cystitis agent in many countries. All of these share the same design principle: hit the bug hard in the urine, spend little of the drug systemically, and keep the course short.

Pyelonephritis — the infection has left the bladder

Once the infection is in kidney tissue or the blood, the drug must achieve good tissue and serum levels — the opposite of the urine-only agents. Treatment is therefore a systemic antibiotic and a longer course (typically 7–14 days rather than 3). Oral options for a stable patient include co-amoxiclav (amoxicillin with the beta-lactamase inhibitor clavulanate, restoring activity against beta-lactamase-producing E. coli), cephalexin (an oral cephalosporin), and ciprofloxacin, a fluoroquinolone with excellent renal-tissue penetration. Fluoroquinolones work superbly here but carry class cautions that have narrowed their use: tendonitis and tendon rupture, aortic aneurysm/dissection, QT prolongation, dysglycaemia, C. difficile risk and neuropsychiatric effects — so they are reserved rather than first-reached-for, and the mechanistic detail (DNA gyrase/topoisomerase inhibition and resistance) lives in the Antimicrobials section. A septic or vomiting patient needs admission and intravenous therapy (for example a broad-spectrum cephalosporin or gentamicin), with local resistance patterns always steering the empirical choice. And any pyelonephritis that fails to settle demands imaging to exclude obstruction or an abscess — pus behind a blockage will never clear on antibiotics alone.

Pregnancy — the same bugs, tighter rules

Two safe drugs, each unsafe at one end of pregnancy — timing is everything. Pregnancy compresses the whole topic into a timing problem. Asymptomatic bacteriuria is actively screened for and treated (unlike in the non-pregnant). But the two first-line cystitis drugs each carry a trimester-specific hazard. Trimethoprim is avoided in the first trimester because folate antagonism raises the risk of neural tube defects; it is the wrong drug in early pregnancy. Nitrofurantoin, conversely, is generally acceptable in the first two trimesters but avoided near term (and in labour) because of the risk of neonatal haemolysis. So the practical teaching is a neat inversion: nitrofurantoin early, trimethoprim (with folate) if needed later — never the reverse. When neither fits, a beta-lactam such as amoxicillin (culture-guided), co-amoxiclav or cephalexin is the reliable pregnancy-safe fallback. This dovetails with the pregnancy-prescribing chapter, which teaches the same trimester-mapping logic across every drug class.

Recurrent UTI — prevention beyond the antibiotic

Recurrent UTI (broadly two or more in six months, or three or more in a year) is managed by climbing a ladder before reaching for continuous antibiotics. Behavioural and simple measures come first: good fluid intake (a randomised trial showed increased water alone cuts recurrences), not delaying voiding, and post-coital voiding where infections are sex-associated. In postmenopausal women, oestrogen deficiency thins the vaginal epithelium and shifts the flora away from protective lactobacilli, raising UTI risk — so topical vaginal oestrogen restores the mucosa and measurably reduces recurrence, a therapy that connects directly to the Menopause and incontinence chapters (the same low-dose local oestrogen used for genitourinary syndrome of menopause). Non-antibiotic options such as methenamine hippurate — which is hydrolysed in acidic urine to formaldehyde, a bladder antiseptic rather than a true antibiotic — are increasingly used to avoid resistance. Only when these fail is antibiotic prophylaxis considered: low-dose nightly nitrofurantoin or trimethoprim, or single post-coital doses, always weighed against the resistance and long-term-toxicity cost of continuous exposure.