Urinary infection is common, mostly straightforward, and the site of two opposite errors that are both widespread: young women with symptoms waiting too long for treatment, and older people without symptoms being treated for a positive dipstick that means nothing.
The single most useful rule on this page. In anyone over 65, and in
anyone with a urinary catheter, a urine dipstick should not be used to diagnose urinary
infection. Bacteria in the urine without symptoms — asymptomatic bacteriuria — is
present in a large proportion of both groups as a normal finding, and it does not need antibiotics.
Treating it does not prevent illness, and it causes resistance, Clostridioides difficile, drug
reactions and falls. The diagnosis in these groups is made on symptoms, not on a strip.
1. Recognising it
| Group | What actually suggests infection |
|---|---|
| Women under 65, not pregnant | Dysuria — burning on passing urine — with new frequency or urgency, and no vaginal discharge or irritation. Two or three of those together make infection likely enough to treat without a test. Suprapubic pain, visible blood, and cloudy urine add weight |
| Men | The same symptoms, but always treated as a complicated infection — longer course, and consider the prostate. Perineal or rectal pain, or fever, suggests prostatitis rather than simple cystitis |
| Pregnancy | Any suspicion is investigated and treated, and asymptomatic bacteriuria is treated in pregnancy — the one place where the rule above reverses, because it raises the risk of pyelonephritis and preterm birth. Send a culture |
| Over 65, or catheterised | New or worsening symptoms referable to the urinary tract — dysuria, new incontinence, suprapubic or flank pain, visible blood, rigors, fever. Not a smell, not cloudiness, and not confusion on its own |
| Children | Often non-specific. Fever without a source, vomiting, poor feeding, lethargy, abdominal pain, new bed-wetting, or offensive urine. Any child under 3 months with a fever needs immediate referral — see babies and children under 5 |
The operational rule, because "do not use a dipstick" is not the same as "do not dip"
- In anyone over 65 or with a catheter, do not send the urine and do not dip it unless there are urinary symptoms. The reason is practical rather than pedantic: once a strip has been done, its result is in the notes, and a positive one is very hard for the next clinician to ignore. The way to avoid acting on a meaningless result is not to generate it.
- "They smell strongly" and "the urine looks cloudy" are not indications to dip. Nor is a fall, nor confusion on its own, nor a routine check in a care home.
- What does justify it: new dysuria, new urinary frequency or urgency, new incontinence, suprapubic or flank pain, visible blood, or rigors — and even then the strip supports the clinical picture rather than replacing it.
- If a strip has already been done and is positive in someone with no urinary symptoms: the correct action is to document that asymptomatic bacteriuria does not require antibiotics and look for another cause of whatever prompted the test. That is a harder conversation than not dipping, which is the argument for not dipping.
- Care homes in particular often dip on a protocol. Anyone in a position to change that protocol is doing more good than any individual prescribing decision.
2. Two things that are not urinary infection, and are treated as one every day
Cloudy or strong-smelling urine, on its own
- Neither predicts infection. Both are usually concentration — a dehydrated person's urine is dark, strong and cloudy, and drinking water fixes it. Phosphate crystals, diet, asparagus, some medicines and vitamins all change smell and appearance.
- Neither is a reason to send urine or start antibiotics in someone with no symptoms. This is the commonest single trigger for an unnecessary prescription in community and care-home practice.
Confusion in an older person with a positive dipstick
- This is the important one, and the reasoning runs the wrong way round. A confused 84-year-old is dipped, the strip is positive because most 84-year-olds' strips are positive, and a diagnosis is made that stops anyone looking further.
- Delirium has a long differential and urinary infection is a minority of it: pneumonia and other infection elsewhere, medication and anticholinergic burden, constipation, urinary retention, pain, dehydration, electrolyte disturbance, alcohol withdrawal, a subdural haematoma from an unremembered fall, hypoxia, and a change of environment.
- Confusion alone, with no urinary symptoms and no signs of systemic infection, is not diagnostic of urinary infection whatever the strip says. Look for a source properly, and treat what you find.
- If they are systemically unwell — fever, rigors, low blood pressure, fast breathing — that is a different question and may be sepsis from a urinary source. Treat urgently. The argument here is against reflex treatment of a strip, not against treating a sick person.
3. Treatment — first choices
Check previous culture results before choosing anything. A past resistant organism changes the answer more than any guideline table does. Send a culture in everyone except a straightforward first episode in a non-pregnant woman under 65.
| Group | First choice | Duration |
|---|---|---|
| Women under 65, not pregnant | Nitrofurantoin, if renal function allows | 3 days |
| Men | Nitrofurantoin or trimethoprim | 7 days |
| Pregnancy | Nitrofurantoin — but not at term. At term, or where nitrofurantoin is unsuitable, cefalexin | 7 days, and send a culture |
| Children over 3 months, lower urinary tract | Trimethoprim or nitrofurantoin, at weight-based dosing | 3 days, and reassess if not improving |
| Catheter-associated, symptomatic | As for the person's group, and consider changing the catheter if it has been in place more than 7 days. Do not treat a positive strip alone | 7 days, or 10 if the response is slow |
| Suspected pyelonephritis — fever, rigors, flank pain | Cefalexin, or co-amoxiclav or ciprofloxacin where culture supports it. Admit if systemically unwell, pregnant, vomiting, or unable to take oral treatment | 7 to 10 days |
The exclusions that decide the choice
- Nitrofurantoin needs working kidneys. Avoid if eGFR is below 45 — it is concentrated in urine and stops working as filtration falls. Between 30 and 44 it may be used with caution, short course, only for a proven or suspected multi-resistant organism where the benefit outweighs the risk. Avoid at term in pregnancy because of neonatal haemolysis, and in known G6PD deficiency. There is MHRA advice on pulmonary and hepatic reactions with long-term use.
- Trimethoprim is contraindicated in the first trimester — it is a folate antagonist. Avoid alongside methotrexate, and take care with potassium-raising drugs and in renal impairment, because it raises potassium. Resistance is common, particularly with use in the last three months and in older people in residential care, so it is a poor blind choice in exactly the group most likely to receive it.
- Amoxicillin only on a sensitive culture. Blind resistance rates are too high for it to be a first choice.
- Quinolones are not first line and carry restrictions on use because of tendon, musculoskeletal and neuropsychiatric adverse effects.
4. Penicillin allergy — which options survive
The trap in this table has a name: pivmecillinam. It is offered as an
alternative when nitrofurantoin and trimethoprim will not do, it is unfamiliar to many prescribers, and
it is a beta-lactam — a penicillin. It must not be used in penicillin allergy, and
because it is neither called "-cillin" in the way amoxicillin is nor recognised on sight, it is the one
most likely to be reached for by mistake.
| Drug | Class | Safe in penicillin allergy? |
|---|---|---|
| Nitrofurantoin | Nitrofuran | Yes — unrelated. Usually the answer |
| Trimethoprim | Antifolate | Yes — unrelated |
| Fosfomycin | Phosphonic acid | Yes — unrelated. Single oral dose, useful where the first two are excluded |
| Pivmecillinam | Beta-lactam (penicillin) | No |
| Amoxicillin, co-amoxiclav | Penicillin | No |
| Cefalexin and other cephalosporins | Beta-lactam, different core | Usually, with a caveat. Cross-reactivity with modern cephalosporins is low — on the order of a few per cent, and lower still for those with dissimilar side chains. Reasonable in a history of rash. Avoid entirely where the reaction was anaphylaxis, angioedema, breathing difficulty, or a severe skin reaction, and seek advice instead |
| Ciprofloxacin | Quinolone | Yes by class, but not a first choice for other reasons |
And ask what the allergy actually was, because most recorded penicillin allergy is not allergy. Childhood rash during a viral illness, nausea, diarrhoea and thrush are all recorded as allergy and none of them is. A label carried for decades narrows every future prescription and is worth unpicking — ask what happened, how soon after the dose, and whether a penicillin has been tolerated since. Referral for formal testing is available and under-used.
5. What to do besides prescribe
- Analgesia works and is under-offered. Paracetamol or ibuprofen for the pain, which is what the person actually came about.
- Fluids, and self-care advice. Many mild episodes in otherwise well women settle without antibiotics, and a shared decision to wait 48 hours with a back-up prescription is a legitimate option.
- Cranberry and D-mannose are not treatments. The evidence for cranberry in prevention of recurrence is weak and inconsistent, and there is none for treating an acute episode. Potassium citrate and sodium bicarbonate preparations relieve symptoms slightly and treat nothing — and alkalinising the urine reduces the effect of nitrofurantoin.
- Safety-net explicitly. Worsening, fever, rigors, flank pain, vomiting or no improvement at 48 hours means being seen again, and the person should be told that in those words.
6. Recurrence, and when to look further
- Investigate rather than re-prescribe where there are three or more episodes a year, any episode in a man or a child, blood in the urine that persists after treatment, suspected stones, or failure to respond to an antibiotic the organism was sensitive to.
- Visible blood in the urine that persists, or any visible blood in someone over 45 without infection, needs urological referral for possible cancer — this is the diagnosis most often lost inside a series of urinary infections.
- In post-menopausal women, vaginal oestrogen reduces recurrence and is the most effective non-antibiotic measure available. It is frequently not offered.
- Consider retention — incomplete emptying is a common and correctable driver in older men and in anyone on anticholinergic medication.
- Antibiotic prophylaxis — single-dose after intercourse, or a nightly low dose — is an option after the above have been addressed, with a plan to review it rather than continue indefinitely.
Where the strip does belong. In a non-pregnant woman under 65 with some but
not all of the classic symptoms, a dipstick genuinely helps — and it helps most as a way of
ruling out, because negative nitrite and negative leukocytes together make infection unlikely.
How to run one properly, how to read it, and what every combination of the ten pads means is covered in
reading a urine dipstick, with an interactive interpreter.