Urinary tract infection is the second most common clinical indication for empirical antimicrobial treatment in primary and secondary care. Urine samples constitute the largest single category of specimens examined in most medical microbiology laboratories. Healthcare practitioners regularly have to make decisions about prescription of antibiotics for urinary tract infection. Criteria for the diagnosis of urinary tract infection vary depending on the patient and the context. Because it is such a common problem it is an excellent opportunity for students to understand how to interpret microbiology results in the context of clinical symptoms and signs.
Normal urine is sterile, so bacteriuria (the presence of bacteria in a urine sample) is abnormal. In domiciliary practice four bacterial species account for 95% of isolates, whereas in hospital there is a wider range of pathogens (Table 20.1).
The clinical significance of bacteriuria depends on how the sample was obtained, how it was transported to the laboratory, and the clinical history. Before considering the clinical significance of bacteriuria in the context of specific problems it is important to understand what the microbiology laboratory means by the term ‘significant bacteriuria’.
Urine sampling and testing
The laboratory confirmation of bacteriuria is made by quantitative culture of an uncontaminated urine specimen. The most reliable methods for obtaining a urine sample without contamination by skin or perineal flora are by suprapubic aspiration of the bladder with a syringe and needle or by urethral catheterization. Neither of these is practical or acceptable in clinical practice so urine samples are collected during micturition. The practice of cleansing the perineum and taking the specimen in the middle of micturition minimizes the risk of contamination by the perineal flora. However, it is critical that cultures of urine are done as quickly as possible, otherwise a few contaminating bacteria from skin and perineal flora will multiply in a few hours at room temperature and give false positive results. Refrigeration and rapid processing in the laboratory can reduce this problem. Alternatively, culture of urine as soon as it is passed can circumvent the possibility of contaminants growing in the urine during transit. This is achieved by using dip-inoculum methods, which consist of agar attached to slides or spoons that are dipped in the urine, drained, and transported in a stoppered bottle to the laboratory where any bacterial colonies are counted after overnight incubation at 37°C. A rough quantification is possible by this method or by other simple methods of direct plating of a fixed volume of urine on appropriate culture media. Simple identification of the resulting significant isolates is made and antimicrobial susceptibility tests performed as described in Chapter 12.
Table 20.1 Bacteria that commonly cause symptomatic urinary tract infection in hospital and general practice and their susceptibility to commonly used oral agents. Note that the bacteria listed account for 95% of all isolates in general practice but only 82% of isolates from hospitals. The remaining 18% of isolates in hospitals are a wide range of bacteria includingEnterococcus spp and Pseudomonas spp.
Criteria of infection
When more than 105 organisms/ml (108/1) of a single bacterial species are cultured from a freshly voided midstream specimen of urine, the term ‘significant’ bacteriuria is used because it is highly unlikely that such a large number of bacteria could be the result of contamination of the sample. In other words the laboratory finding of ‘significant’ bacteriuria means that it is very likely that the bacteria were present in the urine in the body, before micturition.
The symptoms of urinary tract infection can be associated with lower counts of bacteria: 103 organisms/ml in men and 102 organisms/ml in women. However, most clinical laboratories cannot detect such low levels of bacteria with routine methods, so in practice the threshold for significant bacteriuria remains at 105 organisms/ml.
Although normal urine is sterile a few healthy people have bacteriuria without any symptoms. Most information comes from women, in whom the prevalence of bacteriuria increases with age (Fig. 20.1). The prevalence at any given age is related to sexual activity: 5% of married women aged 24-44 are likely to have asymptomatic bacteriuria compared with <1% of nuns of the same age.
There is less information about men. As with women the prevalence of bacteriuria in men increases with age but it is always lower than in women of the same age. In the elderly the prevalence in men and women increases with deterioration of health and reaches 100% in anybody who has an indwelling urinary catheter present for >4 weeks (Table 20.2).
The important point to remember is that bacteriuria is not a disease, it is a laboratory finding (Box 20.1).
Fig. 20.1 Presence of asymptomatic bacteriuria in non-pregnant women by age in six countries.
Table 20.2 Prevalence of bacteriuria in men and women aged 70 years or greater
Urine dipstick testing
Urine dipsticks use reagents to detect the presence of chemicals in the urine. The presence of nitrites in urine is associated with bacteriuria because nitrites are products of bacterial metabolism; however, nitrites can be found in sterile urine as well. The presence of leucocyte esterase is associated with the presence of white blood cells in the urine, which may in turn be associated with urinary tract infection. However, any cause of inflammation of the urinary tract will result in white cells in the urine. Other substances (blood or protein) can be present in the urine of patients with symptomatic urinary tract infection, but they are even less specific to the diagnosis than nitrites or leucocyte esterase.
Box 20.1 Key points about the diagnosis of urinary tract infection
Rigorous analysis of the value of urine dipstick testing has been largely disappointing (Box 20.1). In fact the only patient group in which dipstick testing appears to be helpful is adults with a single symptom (dysuria or frequency). In patients with very clear symptoms of infection the likelihood of bacteriuria is so high that dipstick testing contributes nothing to management. In contrast dipsticks result in too many false negative results to be used for screening of asymptomatic bacteriuria in pregnancy because of the dire consequences of a false negative result.
Common clinical problems
Symptomatic urinary tract infection
Lower urinary tract infection
Infection of the tissues of the bladder or urethra causes increased frequency of micturition with severe burning pain on passing urine (dysuria). The walls of the bladder and urethra have a relatively poor blood supply so bacteria are not able to penetrate into the blood and people with lower urinary tract infection do not have a systemic inflammatory response. In fact if a patient presents with dysuria, frequency and symptoms of systemic inflammatory response it should be assumed that they have upper urinary tract infection.
The symptoms of lower urinary tract infection can be very severe and make it impossible for patients to continue their normal lives without effective treatment.
Cystitis means inflammation of the bladder and is sometimes used as a pseudonym for lower urinary tract infection, although strictly the term should be bacterial cystitis to distinguish from other causes of cystitis (e.g. chemical cystitis or interstitial cystitis).
Upper urinary tract infection
Infection of the kidney causes loin pain and flank tenderness but these are likely to be accompanied by symptoms of lower urinary tract infection as well. The kidney has an excellent blood supply so, in contrast to lower urinary tract infection, patients with upper urinary tract infection commonly have a systemic inflammatory response and may develop bacteraemia with severe sepsis or even septic shock.
Pyelonephritis means inflammation of the kidney and collecting system and is sometimes used as a pseudonym for upper urinary tract infection.
Uncomplicated urinary tract infection
This means lower urinary tract infection in non-pregnant women with no underlying anatomical abnormalities that predispose to recurrent urinary tract infection. The adjective ‘uncomplicated’ is being used in two different ways:
Patients with symptomatic infection of the lower or upper urinary tract benefit from antibiotic treatment. Symptoms would resolve without antibiotic treatment in about 50% of people with lower urinary tract infection but they resolve much faster with antibiotic treatment. Patients with symptoms of upper urinary tract infection require urgent effective antibiotic treatment to minimize the risk of bacteraemia.
In general practice about half the women presenting with frequency and dysuria have sterile urine cultures. This condition is sometimes referred to as the ‘urethral syndrome’ or symptomatic abacteriuria—a common, but largely unexplained condition. Some cases may be due to sexually transmitted organisms, such as chlamydia, and some may represent the early stages of urinary infection. Counselling is more important than antimicrobial therapy, but persistent symptoms need further investigation.
There are only two groups of patients in whom asymptomatic bacteriuria should be treated with antibiotics:
Asymptomatic bacteriuria should not be treated with antibiotics in any other people. Placebo controlled trials have failed to show convincing benefit in patients with diabetes, in institutionalized elderly men or women, or in those with long-term indwelling catheters, whereas the same trials did show increased risk of adverse events in the treated patients, including colonization with antibiotic-resistant bacteria.
Choice of agent
Lower urinary tract infection
In lower urinary tract infection seen in general practice, over 90% of patients become asymptomatic after a few days' appropriate antibiotic therapy and remain free from bacteriuria for several weeks or more. Most current practice guidelines recommend empirical treatment with a ‘best guess’ antibiotic selected based on knowledge of likely pathogens and local resistance patterns. In domiciliary practice Escherichia colipredominates and most will be fully sensitive to all the commonly used antimicrobials listed in Table 20.1. However, ampicillin-resistant organisms are now sufficiently common for this drug to be abandoned in favour of trimethoprim or one of the other oral agents. The use of co-trimoxazole is not recommended in the treatment of urinary infection, since the sulphonamide component plays an insignificant role and trimethoprim alone is less toxic.
Two agents, nitrofurantoin and nalidixic acid, achieve adequate concentrations only in urine and are exclusively used in lower urinary tract infection. Nitrofurantoin has the distinct advantage of being unrelated to other antibiotics. In contrast nalidixic acid is a quinolone and is likely to select for bacteria that are resistant to fluoroquinolones such as ciprofloxacin. For that reason nitrofurantoin is usually recommended as the preferred alternative to trimethoprim for lower urinary tract infection.
For bacteria that are resistant to nitrofurantoin and trimethoprim or for patients who cannot tolerate these drugs there is a range of alternative oral agents, including oral cephalosporins, co-amoxiclav, fluoroquinolones, and pivmecillinam. The reason that these drugs are not used first line is that they are no more effective than nitrofurantoin or trimethoprim and should be reserved for patients in whom these first line agents are ineffective or contraindicated.
Upper urinary tract infection
Upper urinary tract infection can be accompanied by bacteraemia, making it a life-threatening infection. However, infection should respond to effective antibiotic treatment in 90% of patients.
Nitrofurantoin is not an effective treatment for upper urinary tract infection because it does not achieve effective concentrations in the blood. Resistance to trimethoprim is too common to recommend this drug for empirical treatment of a life-threatening infection. Consequently, empirical treatment should be with a broad-spectrum antibiotic such as co-amoxiclav or ciprofloxacin.
Because of the potentially serious consequences of upper urinary tract infections it is recommended that urine cultures should be obtained before starting antibiotic treatment in all patients. This is because community-acquired infection can be caused by pathogens that are resistant to either co-amoxiclav, ciprofloxacin or any of the other oral antibiotics used in general practice and this is not an acceptable risk with a life-threatening infection.
In the special circumstances of urinary infection, unlike those in other parts of the body, the drugs used are often preferentially excreted into the urine and may attain very high concentrations there, sometimes for long periods. Moreover, in the treatment of lower urinary tract infection (in contrast to pyelonephritis or infections complicating urinary tract abnormalities) antibacterial drugs are generally needed only to tip the balance in favour of normal clearance mechanisms. Several studies have shown that muchcurtailed regimens, lasting 1-3 days, are as successful as prolonged therapy in curing acute urinary infections. Indeed, longer courses are wasteful of resources especially since many patients, wiser than their doctors, abandon treatment once the symptoms abate.
Short-course treatment has an additional potential benefit in serving to identify those few patients (the ones who fail short-course therapy) who are likely to require more extensive urological investigation. Most current guidelines recommend 3 days' treatment with trimethoprim for uncomplicated lower urinary tract infection but there is less certainty about nitrofurantoin. UK guidelines recommend 3 days' therapy but guidelines in other countries recommend treatment for 5 or 7 days.
Management of common clinical problems
Acute symptomatic infections in children
Diagnosis of symptomatic urinary tract infection is rarely straightforward, especially in younger children. They may present with generalized symptoms (fever, vomiting, general malaise) rather than with symptoms in the urinary tract. Consequently, clinically, suspicion should be confirmed by urine culture. If it is difficult to obtain a high-quality clean catch midstream specimen of urine then diagnosis may have to rely on obtaining urine by catheterization or suprapubic needle aspirate. In a child with a low clinical suspicion of urinary tract infection in whom these tests are considered unnecessarily invasive urine dipstick testing can be used, followed by culture of urine only if the dipstick results suggest bacteriuria. However, false negative dipstick tests do occur.
Acute symptomatic infections in adult women
Clinical diagnosis of lower urinary tract infection is reliable in young adult women who have dysuria and frequency but no history of vaginal discharge. Neither dipstick tests nor urine culture are necessary to confirm the diagnosis and empiric antibiotic treatment (3 days of trimethoprim or nitrofurantoin) should be given on the basis of these symptoms alone. If the symptoms are less clear-cut (e.g. the patient has frequency or dysuria but not both) then urine dipstick testing should be done. If this is positive then 3 days of trimethoprim or nitrofurantoin should be given, but if it is negative bacteriuria should be confirmed with culture before treatment is given.
If the woman has symptoms of upper tract infection (loin pain or systemic inflammatory response) then a urine culture should be taken before empiric treatment is started in order to identify resistant bacteria. Empiric treatment should be with co-amoxiclav, pivmecillinam, or a fluoroquinolone for 7 days.
Urinary tract infection is difficult to diagnose in older women because it is more likely to present with vague, generalized symptoms. Moreover the prevalence of asymptomatic bacteriuria increases steadily with age and with increasing co-morbidity. Over 50% of institutionalized elderly women have asymptomatic bacteriuria all of the time. The decision to give antibiotic treatment should be based on clinical diagnosis of infection, supported by acute local or systemic symptoms of inflammation. Smelly urine just means that the patient has bacteriuria, which is not unusual and does not require antibiotic treatment.
Recurrent symptomatic infections in women
Recurrent urinary tract infection in healthy non-pregnant women is defined as three or more episodes during a 12-month period. Antibiotics can be used to reduce the frequency of recurrent infection in two ways. A single dose of either trimethoprim or nitrofurantoin taken at night reduces the risk of symptomatic infection to about one-fifth of the risk with no treatment. However, the risk of recurrent urinary infection returns to pretreatment levels as soon as treatment is stopped. An alternative, equally effective strategy for women with infection associated with sexual intercourse is to take a single postcoital dose of antibiotic. Prophylactic antibiotics for recurrent infection have side effects (oral or vaginal candidiasis and gastrointestinal symptoms) but infection by bacteria resistant to the prophylactic antibiotic does not appear to be a significant risk.
In postmenopausal women oestrogen replacement is not consistently effective in preventing recurrent urinary tract infection and is less effective than antibiotic prophylaxis.
Cranberry products (juice and capsules or tablets containing concentrated extracts) are effective for preventing recurrent infection in young women after antibiotic treatment of an acute attack. However, there is no evidence about their effectiveness in post-menopausal women or in comparison with antibiotic prophylaxis.
Acute symptomatic infections in men
Uncomplicated lower urinary tract infection does not occur in men. Urinary tract infections in men are generally viewed as complicated because they result from an anatomic or functional anomaly or instrumentation of the genito-urinary tract. Consequently, urine cultures should be obtained before antibiotic treatment is started. It is impossible to distinguish reliably between urinary tract infection and prostatitis, consequently 2 weeks' empirical treatment with a fluoroquinolone is recommended.
Between 2% and 7% of patients with indwelling urethral catheters acquire bacteriuria each day, even with the application of best practice for insertion and care of the catheter. All patients with a long-term indwelling catheter are bacteriuric, often with two or more organisms. The presence of a short- or long-term indwelling catheter is associated with a greater incidence of fever of urinary tract origin. Fever without any localizing signs is a common occurrence in catheterized patients and urinary tract infection accounts for about a third of these episodes.
In catheterized patients the common occurrence of fever, the consistent presence of bacteriuria, and the variable presence of a broad range of other associated clinical manifestations (new onset confusion, renal angle tenderness or suprapubic pain, chills, rigors, etc.) makes the diagnosis of symptomatic urinary tract infection difficult.
The presence of one of the following symptoms is an indication for empiric antibiotic treatment:
Urine culture should be used to test the susceptibility of the bacteria that are inevitably present in the urine. Antibiotic treatment should never be given simply because of change in the smell or appearance of the urine in patients with indwelling urinary catheters.
Albert X, Huertas I, Pereiró I, Sanfélix J, Gosalbes V, Perrota C. Antibiotics for preventing recurrent urinary tract infection in non-pregnant women. The Cochrane Database of Systematic Reviews 2004, Issue 3. Art. No.: CD001209. DOI: 10.1002/14651858.CD001209.pub2.
Cincinnati Children's Hospital Medical Center. Evidence based clinical practice guideline for medical management of first time acute urinary tract infection in children 12 years of age or less. Cincinnati (OH): Cincinnati Children's Hospital Medical Center; 2005. Available at: http://www.guideline.gov/summary/summary.aspx?doc_id=7272&nbr=004334&string=uti+AND+children
Scottish Intercollegiate Guidelines Network. Management of suspected bacterial urinary tract infection in adults. SIGN, Edinburgh, 2006. Available at: http://www.sign.ac.uk/pdf/sign88.pdf