Section 4 Clinical manifestation and diagnosis of infections by body system
20 Urinary tract infections
Urinary tract infections are common, especially among women
The urinary tract is one of the most common sites of bacterial infection, particularly in females; 20–30% of women have recurrent urinary tract infections (UTIs) at some time in their life. UTIs in men are less common and primarily occur after 50 years of age. Although the majority of infections are acute and short-lived, they contribute to a significant amount of morbidity in the population. Severe infections result in a loss of renal function and serious long-term sequelae. In females, a distinction is made between cystitis, urethritis and vaginitis, but the genitourinary tract is a continuum and the symptoms often overlap.
Acquisition and etiology
Bacterial infection is usually acquired by the ascending route from the urethra to the bladder
The infection may then proceed to the kidney. Occasionally, bacteria infecting the urinary tract invade the bloodstream to cause septicaemia. Less commonly, infection may result from haematogenous spread of an organism to the kidney, with the renal tissue being the first part of the tract to be infected.
From an epidemiological viewpoint, UTIs occur in two general settings: community-acquired and hospital (nosocomially)-acquired, the latter most often being associated with catheterization. Hospital-acquired UTIs, while less common than community acquired, contribute significantly (ca. 40%) to overall nosocomial infection rates.
The Gram-negative rod Escherichia coli is the commonest cause of ascending UTI
Other members of the Enterobacteriaceae are also implicated (Fig. 20.1). Proteus mirabilis is often associated with urinary stones (calculi), probably because this organism produces a potent urease, which acts on urea to produce ammonia, rendering the urine alkaline. Citrobacter, Klebsiella,Enterobacter, Proteus, and Pseudomonas aeruginosa are more frequently found in hospital-acquired UTI because their resistance to antibiotics favours their selection in hospital patients (see Ch. 36).
Figure 20.1 Common causes of urinary tract infection. The percentages of infections caused by different bacteria in outpatients and hospital inpatients are shown. E. coli is by far the commonest isolate in both groups of patients, but note the difference in the percentage of infections caused by other Gram-negative rods. These isolates often carry multiple antibiotic resistance and colonize patients in hospital, especially those receiving antibiotics.
Among the Gram-positive species, Staphylococcus saprophyticus has a particular propensity for causing infections, especially in young sexually active women. Staphylococcus epidermidis and Enterococcus species are more often associated with UTI in hospitalized patients (especially those with AIDS), where multiple antibiotic resistance can cause treatment difficulties. In some instances, capnophilic species (organisms that grow better in air enriched with carbon dioxide), including corynebacteria and lactobacilli, have been implicated as possible causes of UTI. Obligate anaerobes are very rarely involved.
When there has been haematogenous spread to the urinary tract, other species may be found, e.g. Salmonella typhi, Staphylococcus aureus and Mycobacterium tuberculosis (renal tuberculosis).
Viral causes of UTI appear to be rare, although there are associations with haemorrhagic cystitis and other renal syndromes
Certain viruses may be recovered from the urine in the absence of urinary tract disease and include:
• The human polyomaviruses, JC and BK, enter the body via the respiratory tract, spread through the body and infect epithelial cells in the kidney tubules and ureter, where they establish latency with persistence of the viral genome. About 35% of kidneys from healthy individuals contain polyomavirus DNA sequences. However, during normal pregnancy, the viruses may reactivate asymptomatically, with the appearance of large amounts of virus in the urine. Reactivation also occurs in immunocompromised patients (see Ch. 30) and may lead to haemorrhagic cystitis.
• High titres of cytomegalovirus (CMV) and rubella may be shed asymptomatically in the urine of congenitally infected infants (see Ch. 23).
• In contrast to asymptomatic shedding, some serotypes of adenovirus have been implicated as a cause of haemorrhagic cystitis.
• The rodent-borne hantavirus responsible for Korean haemorrhagic fever infects capillary blood vessels in the kidney and can cause a renal syndrome with proteinuria.
• Finally, a number of other viruses can infect the kidneys, including mumps and HIV.
Urine samples are commonly investigated by virus isolation, immunological and genomic detection methods.
Very few parasites cause UTIs
Other causes of UTI include:
• The fungi Candida spp. and Histoplasma capsulatum.
• The protozoan Trichomonas vaginalis (see Ch. 21), which can cause urethritis in both males and females, but is most often considered as a cause of vaginitis.
• Infections with Schistosoma haematobium (see Ch. 27), which result in inflammation of the bladder and commonly haematuria. The eggs penetrate the bladder wall, and in severe infections large granulomatous reactions can occur and the eggs may become calcified. Bladder cancer is associated with chronic infections, although the mechanism is uncertain. Obstruction of the ureter as a result of egg-induced inflammatory changes can also lead to hydronephrosis.
A variety of mechanical factors predispose to UTI
Anything that disrupts normal urine flow or complete emptying of the bladder or facilitates access of organisms to the bladder will predispose an individual to infection (Fig. 20.2). The shorter female urethra is a less effective deterrent to infection than the male urethra (see Ch. 13). Sexual intercourse facilitates the movement of organisms up the urethra, particularly in females, so the incidence of UTI is higher among sexually active women than among celibate women. Preceding bacterial colonization of the periurethral area of the vagina is perhaps important (see below).
Figure 20.2 Bacterial attributes and host factors favouring urinary tract infection (UTI). Abnormalities of the urinary tract tend to predispose to infection. Bacterial adherence factors have been studied in detail, but relatively little is known about other bacterial virulence factors in UTI.
In male infants, UTIs are more common in the uncircumcised, and this is associated with colonization of the inside of the prepuce and urethra with faecal organisms.
Pregnancy, prostatic hypertrophy, renal calculi, tumours and strictures are the main causes of obstruction to complete bladder emptying
Increased volumes of post-void residual urine are associated with a greater likelihood of infection. Infection, superimposed on urinary tract obstruction, may lead to ascent of infection to the kidney and rapid destruction of renal tissue.
Loss of neurologic control of the bladder and sphincters (e.g. in spina bifida, paraplegia or multiple sclerosis), and the resultant large residual volume of urine in the bladder, causes a functional obstruction to urine flow, and such patients are particularly prone to recurrent infections.
Vesicoureteral reflux (reflux of urine from the bladder cavity up the ureters, sometimes into the renal pelvis or parenchyma) is common in children with anatomic abnormalities of the urinary tract and may predispose to ascending infection and kidney damage. Reflux may also occur in association with infection in children without underlying abnormalities, but tends to disappear with age.
Clinical studies including reports that pyelonephritis (infection of the kidney) is commonly found in people with diabetes mellitus at post-mortem suggest an increased propensity for UTI in individuals with diabetes mellitus. People with diabetes mellitus may have more severe UTIs, and if diabetic neuropathy interferes with normal bladder function, persistent UTIs are common.
Catheterization is a major predisposing factor for UTI
During insertion of the catheter, bacteria may be carried directly into the bladder and, while in situ, the catheter facilitates bacterial access to the bladder either via the lumen of the catheter or by tracking up between the outside of the catheter and the urethral wall (Fig. 20.3). The catheter disrupts the normal bladder’s protective function action and allows bacteria to get a foothold. Thus, duration of catheterization is directly associated with increased probability of infection (i.e. risk of UTI increases by about 3–10% each day of catheterization).
Figure 20.3 The urinary catheter. Catheterization is an important predisposing factor for infection. Bacteria can be pushed into the bladder as the catheter is inserted and, while the catheter is in place, bacteria reach the bladder by tracking up between the outside of the catheter and the urethra. Contamination of the catheter drainage system by bacteria from other sources can also result in infection. Specimens of bladder urine for laboratory investigations can be collected from catheterized patients as shown. The second port (above) is for putting fluids into the bladder. Urine from the drainage bag should not be tested because it may have been standing for several hours.
A variety of virulence factors are present in the causative organisms
The conflict between host and parasite in the urinary tract has been discussed in Chapter 13. Most urinary tract pathogens originate in the faecal flora, but only the aerobic and facultative species such as E. coli possess the attributes required to colonize and infect the urinary tract. The ability to cause infection of the urinary tract is limited to certain serogroups of E. coli such as O (semantic) serotypes (e.g. O1, O2, O4, O6, O7 and O75) and K (capsular) serotypes (e.g. K1, K2, K3, K5, K12 and K13). These serotypes differ from those associated with gastrointestinal tract infection (see Ch. 22), which has led to use of the term ‘uropathogenic E. coli’ (UPEC). The success of these strains is attributable to a variety of genes in chromosomal pathogenicity islands (see Ch. 2) which are not found in faecal E. coli. For example, UPEC typically contains genes associated with colonization of the periurethral areas. A prime example is the adhesion known as P. fimbriae (pyelonephritis-associated pili (PAP)), which allows UPEC to specifically adhere to urethral and bladder epithelium. Studies with other species of urinary tract pathogens have confirmed the presence of similar adhesins for uroepithelial cells (Fig. 20.4).
Figure 20.4 Scanning electron micrograph showing bacteria attached to an exfoliated uroepithelial cell from a patient with acute cystitis.
(Courtesy of T.S.J. Elliot and the editor of British Journal of Urology.)
Other features of E. coli which appear to assist in the localization of organisms in the kidney and in renal damage include:
• The capsular acid polysaccharide (K) antigens are associated with the ability to cause pyelonephritis and are known to enable E. coli strains to resist host defences by inhibiting phagocytosis.
• Haemolysin production by E. coli is linked with the capacity to cause kidney damage; many haemolysins act more generally as membrane-damaging toxins.
The production of urease by organisms such as Proteus spp. has been correlated with their ability to cause pyelonephritis and stones.
The healthy urinary tract is resistant to bacterial colonization
With the exception of the urethral mucosa, the urinary tract usually eliminates microorganisms rapidly and efficiently (see Ch. 13). The pH, chemical content and flushing mechanism of urine help to dispose of organisms in the urethra. Although urine is a good culture medium for most bacteria, it is inhibitory to some, and anaerobes and other species (non-haemolytic streptococci, corynebacteria and staphylococci), which comprise most of the normal urethral flora, do not readily multiply in urine.
Although the inflammatory response to urinary tract infection involves leukocytic, chemokine, and cytokine response, the role of humoral immunity in the host’s defence against infection of the urinary tract is poorly understood. After infection of the kidney, IgG and secretory IgA antibodies can be detected in urine but do not appear to protect against subsequent infection. Infection of the lower urinary tract is usually associated with a low or undetectable serologic response, reflecting the superficial nature of the infection; the bladder and urethral mucosa are rarely invaded in UTIs.
Clinical features and complications
Acute lower UTIs cause dysuria, urgency and frequency
Acute infections of the lower urinary tract are characterized by a rapid onset of:
• dysuria (burning pain on passing urine)
• urgency (the urgent need to pass urine)
• frequency of micturition.
However, UTIs in the elderly and those with indwelling catheters are usually asymptomatic.
The urine is cloudy due to the presence of pus cells (pyuria) and bacteria (bacteriuria), and may contain blood (haematuria). Examination of urine specimens in the laboratory is essential to confirm the diagnosis. Patients with genital tract infections such as vaginal thrush or chlamydial urethritis may present with similar symptoms (see Ch. 21).
Pyuria in the absence of positive urine cultures can be due to chlamydiae or tuberculosis and is also seen in patients receiving antibacterial therapy for UTI, as the bacteria are inhibited or killed by the antibacterial agent before the inflammatory response dies away.
Recurrent infections of the lower urinary tract occur in a significant proportion of patients. They may be:
• relapses, caused by the same strain of organism
• reinfections by different organisms.
Recurrent infections can result in chronic inflammatory changes in the bladder, prostate and periurethral glands.
Acute bacterial prostatitis causes systemic symptoms (fever) and local symptoms (perineal and low back pain, dysuria and frequency)
Acute bacterial prostatitis may arise from ascending or haematogenous infection, and people lacking the antibacterial substances normally present in prostatic fluid are perhaps more susceptible. Chronic bacterial prostatitis, however, although usually caused by E. coli, is difficult to cure and can be a source of relapsing infection within the urinary tract.
Although it may be important to know whether an infection is restricted to the bladder (lower urinary tract) or has ascended to the upper urinary tract and kidney, there are no satisfactory methods for distinguishing the two other than by examining urine directly from the ureter by ureteric catheterization.
Pyelonephritis causes a fever and lower urinary tract symptoms
Patients with pyelonephritis (infection of the kidney, Fig. 20.5) present with lower urinary tract symptoms and usually have a fever. Staphylococci are a common cause and renal abscesses are generally present. Recurrent episodes of pyelonephritis result in a loss of function of renal tissue, which may in turn cause hypertension, itself a cause of renal damage. Infection associated with stone formation can result in obstruction of the renal tract and septicaemia.
Figure 20.5 Histologic appearance of the kidney in acute pyelonephritis, showing the intense inflammatory reaction and microabscesses (M). (H&E stain.)
(Courtesy of M.J. Wood.)
Haematuria is a feature of endocarditis and a manifestation of immune complex disease, as well as a result of infections of the kidney, and its presence warrants careful investigation. Pyuria may be associated with kidney infection by M. tuberculosis. This organism cannot be grown by normal urine culture methods and therefore the patient may appear to have a sterile pyuria.
Asymptomatic infection (i.e. significant numbers of bacteria in the urine in the absence of symptoms, see below) can be detected only by screening urine samples in the laboratory. It is important in instances such as:
• pregnant women and young children, where failure to treat may result in chronic renal damage
• people undergoing instrumentation of the urinary tract, in whom bacteriuria may proceed to bacteraemia
• the elderly and those with diabetes (both risk factors for asymptomatic bacteriuria).
A key feature is the detection of significant bacteriuria.
Infection can be distinguished from contamination by quantitative culture methods
In health, the urinary tract is sterile, although the distal region of the urethra is colonized with commensal organisms, which may include periurethral and faecal organisms. As urine specimens are usually collected by voiding a specimen into a sterile container, they become contaminated with the periurethral flora during collection. Infection can be distinguished from contamination by quantitative culture methods. Bacteriuria is defined as ‘significant’ when a properly collected midstream urine (MSU) specimen is shown to contain over 105 organisms/mL. Infected urine usually contains only a single bacterial species. Contaminated urine usually has < 104 organisms/mL and often contains more than one bacterial species (Fig. 20.6). Distinguishing infection from contamination when counts are 104–105 organisms/mL can be difficult. Careful collection and rapid transport of urine specimens to the laboratory are essential (see below and Ch. 32).
Figure 20.6 Significant bacteriuria. Voided specimens of urine are rarely sterile because the urine is contaminated with organisms from the periurethral area during collection. Even well-collected specimens from healthy individuals may contain up to 103 bacteria/mL of urine. A count of 105 bacteria/mL is considered a reliable indicator of infection. However, there are various reasons why lower counts may sometimes be significant (e.g. acute dysuria, ureteral obstruction, etc.).
It is important to recognize that the criteria for ‘significant bacteriuria’ do not apply to urine specimens collected from catheters or nephrostomy tubes or by suprapubic aspiration directly from the bladder, in which any number of organisms may be significant because the specimen is not contaminated by periurethral flora. In addition, infection of sites in the urinary tract below the bladder, and by organisms that are not members of the normal faecal flora, may not lead to the presence of significant numbers in the urine.
The usual urine specimen for microbiologic examination is an MSU sample
An MSU sample should be collected into a sterile wide-mouthed container after careful cleansing of the labia or glans with soap (not antiseptic) and water, and after allowing the first part of the urine stream to be voided, as this helps to wash out contaminants in the lower urethra. After suitable instruction, the majority of adult patients can collect satisfactory samples with minimum supervision, though collection may be difficult for elderly and bedridden patients and consideration should be given to these difficulties when interpreting results.
Collection of MSU samples from babies and young children is difficult. ‘Bag urine’ may be collected by sticking a plastic bag to the perineum in girls or to the penis in boys, but such specimens are frequently heavily contaminated with faecal organisms. These problems can be overcome by suprapubic aspiration of urine directly from the bladder (Fig. 20.7).
Figure 20.7 Suprapubic aspiration of bladder urine. Urine samples can be collected directly from the bladder by insertion of a needle. This method is useful in young children from whom it is difficult to obtain uncontaminated midstream urine specimens.
Urine specimens should be transported to the laboratory with minimum delay because urine is a good growth medium for many bacteria and multiplication of organisms in the specimen between collection and culture will distort the results (see Ch. 32).
Ideally, samples should be collected before antimicrobial therapy is started. If the patient is receiving, or has received, therapy within the previous 48 h, this should be stated clearly on the request form.
For patients with a catheter, a catheter specimen of urine is used for microbiologic examination
Patients should not be catheterized simply to obtain a urine sample. Urine is obtained from patients who have a catheter in situ by withdrawing a sample with a syringe and needle from the catheter tube as shown in Figure 20.3. Urine that has been standing in the catheter drainage bag for hours is unsuitable for testing because the organisms may have multiplied to give much greater numbers than those present in the patient.
Special urine samples are required to detect M. tuberculosis and Schistosoma haematobium
• three early morning urine samples on consecutive days for M. tuberculosis; these do not require the same precautions during collection as an MSU sample, because the culture technique prohibits the growth of organisms other than mycobacteria
• the last few millilitres of a urine sample collected early afternoon after exercise for detection of S. haematobium.
Urine specimens should be examined macroscopically and microscopically and should be cultured by quantitative or semiquantitative methods, as summarized in Chapter 32.
Microscopic examination of urine allows a rapid preliminary report
Bacteria may be seen on microscopy when present in the specimen in large numbers. However, they are not necessarily indicative of infection, but may indicate that the specimen has been poorly collected or left at room temperature for a prolonged period of time.
The presence of red and white blood cells, although abnormal, is not necessarily indicative of UTI. Haematuria may be present in association with:
• infection of the urinary tract and elsewhere (e.g. bacterial endocarditis)
• renal trauma
• urinary tract carcinomas
• clotting disorders
Occasionally, red blood cells may contaminate urine specimens of menstruating women.
White blood cells are present in the urine in very small numbers (e.g. < 10/mL) in health; a count of over 10/mL is considered abnormal, but is not always associated with bacteriuria. Sterile pyuria is an important finding and may reflect:
• concurrent antibiotic therapy
• other diseases such as neoplasms or urinary calculi
• infection with organisms not detected by routine urine culture methods.
Renal tubular cells, seen in the urine of aspirin-misusers, may be confused with white blood cells. Urinary casts are also indicative of renal tubular damage.
A laboratory diagnosis of significant bacteriuria requires quantification of the bacteria
Conventional culture methods produce results within 18–24 h, but rapid methods (e.g. based on bioluminescence, turbidimetry, leukocyte esterase/nitrate reductase test, etc.) are also available. In some laboratories, direct antibiotic susceptibility tests may be initiated upon detecting abnormal numbers of white blood cells or bacteria on microscopy, so that both culture and susceptibility results are available within 24 h.
Interpretation of the significance of bacterial culture results depends upon a variety of factors
These factors relate to:
• collection – specimen collection must be carried out properly.
• storage – the urine must be cultured within 1 h of collection or held at 4°C for not more than 18 h before culture.
• antibiotic treatment – in a patient receiving antibiotics, smaller numbers of organisms may be significant and may represent an emerging resistant population; simple laboratory methods are available to detect antibacterial substances.
• fluid intake – the patient may be taking more or less fluid than usual, and this will clearly influence the quantitative result.
• the specimen – the quantitative guidelines are valid for MSU specimens; they do not apply to catheter specimens, suprapubic aspirates or nephrostomy samples.
Depending on clinical evaluation of the patient and local antimicrobial resistance trends, uncomplicated UTI is typically treated with an oral antibacterial for 3 days
Uncomplicated UTI (cystitis) generally resolves spontaneously within 4 weeks in up to 40% of patients; however, treatment with antibacterial agents reduces symptoms and ensures bacterial eradication. Oral antimicrobial chemotherapy is generally administered twice a day for 3 days, depending upon the drug and clinical evaluation of the patient. The commonly prescribed agents are shown in Table 20.1. The choice of agent should be based on the results of susceptibility tests. However, for uncomplicated UTIs in patients in the community, therapy is often ‘best guess’, at least until laboratory results are available. This requires a knowledge of the likely pathogens and their antibiotic susceptibility patterns in the locality. Follow-up cultures should be carried out after treatment has been completed (at least 2 days later) to confirm eradication of the infecting organism. In addition to antibacterial therapy, the patient should be advised to drink large volumes of fluid to help the normal flushing process.
Table 20.1 Oral antibacterials for urinary tract infections (UTIs)
Examples of common oral antibacterials for urinary tract infections
Class of agent
Antimetabolite/nucleic-acid synthesis inhibitor
Incidence of resistant strains increasing
Combination of trimethoprim with sulphamethoxazole (also antimetabolite nucleic-acid synthesis inhibitor)
One of the most common ‘first-line’ therapeutic approaches; may be useful in ‘blind’ treatment but more toxic than trimethoprim alone; resistance also an issue
For uncomplicated UTI caused by E. coli and Staphylococcus saprophyticus; not active in alkaline pH (therefore not useful for Proteus infections)
Ciprofloxacin, norfloxacin, ofloxacin, etc.
Very broad spectrum; not highly active against enterococci; increasing resistance an issue
Several different classes of antibacterial are available in oral formulations and suitable for treatment of UTI. Nitrofurantoin is useful only for lower UTIs as it does not achieve adequate serum and tissue concentrations to treat upper UTIs.
Children and pregnant women with asymptomatic bacteriuria should be treated with antibacterials and followed-up to check for eradication of the infection. Instrumentation of the urinary tract should be delayed in patients with significant bacteriuria until appropriate treatment has rendered the urine sterile.
Complicated UTI (pyelonephritis) should be treated with a systemic antibacterial agent
The organism should be known to be susceptible to the antibacterial, and systemic treatment should continue until the signs and symptoms subside. It can then be replaced by oral therapy. The usual length of treatment is at least 10 days, but longer treatment may be necessary to sterilize the kidney.
Hospital-acquired infections or recurrent infections, particularly in catheterized patients, may be caused by antibiotic-resistant organisms, and the agent of choice will depend upon the antibacterial susceptibility pattern. If possible, the catheter should be removed, as eradication of infection is extremely difficult to achieve in catheterized patients, and some would advocate treatment only when the patient complains of symptoms or before invasive procedures. Guidelines for catheter care and for the prevention of catheter-associated UTIs are shown in Box 20.1.
Box 20.1 Guidelines for catheter care
• Avoid catheterization whenever possible.
• Keep duration of catheterization to a minimum.
• Use intermittent rather than continuous catheterization when feasible.
• Insert catheters with good aseptic technique.
• Use a closed sterile drainage system.
• Maintain a gravity drain.
• Use topical antiseptics around the meatus in women.
• Wash hands before and after inserting catheters and collecting specimens, and after emptying drainage bags.
• Catheters that drain into open collecting vessels are highly conducive to infection. Thus, closed drainage systems are now used in most hospitals, but even then, bacteriuria occurs in a significant number of patients.
Infections acquired by haematogenous spread require specific antibacterial therapy, as described in Chapter 33 for tuberculosis, Chapter 22 for S. typhi, Chapter 26 for Staph. aureus and Chapter 27 for schistosomiasis.
Many of the features of the pathogenesis of UTI and host predispositions are not clearly understood
Recurrent infections in otherwise healthy women can be prevented by regularly emptying the bladder. This washes bacteria out of the urinary tract and is particularly important following intercourse. The prophylactic use of antibiotics may also prevent recurrent infections, but in the presence of underlying abnormalities, there is a tendency to select antibiotic-resistant strains, which subsequently cause infections that are more difficult to treat.
Infection in catheterized patients is very common, but can be reduced by good catheter care procedures (Box 20.1, see Ch. 36). Catheterization should be avoided if possible or kept to a minimum duration.
• UTIs are among the commonest bacterial infections, especially in women.
• Most UTIs are acute episodes without sequelae.
• UTIs are usually endogenously acquired, with colonizing bacteria ascending the urinary tract from the periurethral area. E. coli is the predominant pathogen; other Gram-negative rods are also responsible, especially in hospitalized patients. Viruses are not important causes of UTI.
• Structural or mechanical factors in the host, or catheterization, predispose to infection.
• Bacterial attributes such as adhesions and capsular polysaccharides may be important in the development of UTI. Specific toxins are not implicated, but haemolysins (cytotoxins) may be.
• Lower UTI usually presents with acute frequency and dysuria. Asymptomatic infection is common in pregnancy and in children. Infection is recurrent in a significant proportion of people.
• Pyelonephritis (upper UTI) has a more severe presentation than lower UTI, with fever and loin pain; recurrent infection results in renal damage.
• Bacteriologic confirmation of the diagnosis requires quantitative methods. Pyuria also implies infection.
• Short-course treatment with oral antibacterials is effective for lower UTI; pyelonephritis needs longer treatment, often commencing with systemically administered drugs.
• Hospital-acquired UTI is often caused by multiple-resistant Gram-negative bacteria, and treatment should be based on the results of antibiotic susceptibility tests.