Common use of Patient characteristics Clause in Contracts

Patient characteristics. Two comparisons were made for the purposes of this inves- tigation: (1) patients with bacteriuria vs nonbacteriuria control subjects (Table 1), and (2) UTI vs ASB subjects (Table 2). All 296 study subjects were white. ■ Patients with bacteriuria vs control subjects —When com- pared with individuals from the nonbacteriuria control group ( Table 1), patients with bacteriuria (either UTI or ASB) were more often: ▫ female (84% vs 75%; OR, 1.84; 95% CI, 1.03-3.26) ▫ nursing home residents (OR, 1.74; 95% CI, 1.08-2.80) ▫ had a history of hypertension (OR, 1.81; 95% CI, 1.14-2.87) ▫ received bladder catheterization in the ED or on admission (OR, 1.72; 95% CI, 1.08-2.74) JAOA • Vol 109 • No 4 • April 2009 • 221 ■ Mean (SD) Age, y 83 (8) 82 (8) ... ■ Female 130 (84) 106 (75) 1.84 (1.03-3.26)† ■ Nursing Home Resident 68 (44) 47 (33) 1.74 (1.08-2.80)† ■ Medical History ▫ Cognitive impairment 47 (31) 38 (27) 1.20 (0.73-1.99) ▫ Cardiovascular health – Hypertension 93 (60) 65 (46) 1.81 (1.14-2.87)† – Stroke 30 (19) 34 (24) 0.77 (0.44-1.34) ▫ Pulmonary disease – Chronic obstructive pulmonary disease 33 (21) 36 (25) 0.80 (0.47-1.38) – Pneumonia 16 (10) 19 (13) 0.75 (0.37-1.52) ▫ Diabetes mellitus 45 (29) 34 (24) 1.31 (0.78-2.20) ▫ Falls 16 (10) 9 (6) 1.71 (0.73-4.01) ■ Hospital Care‡ ▫ Bladder catheter 78 (51) 53 (37) 1.72 (1.08-2.74)† ▫ Antibiotic treatment 92 (60) 67 (47) 1.66 (1.05-2.63) † – Levofloxacin treament 53 (58) 26 (39) 2.09 (1.10-3.96)† ■ Ischuria§ 5 (4) 11 (8) 0.44 (0.15-1.31) ■ Delirium 46 (30) 11 (8) 5.07 (2.51-10.27)† ■ Cardiovascular Health ▫ Atrial fibrillation 19 (12) 19 (13) 0.91 (0.46-1.80) ▫ Congestive heart failure 31 (20) 18 (13) 1.74 (0.92-3.27) ■ Pulmonary Disease 42 (27) 59 (42) 0.53 (0.32-0.86)† ▫ Pneumonia 23 (15) 32 (23) 0.60 (0.33-1.09) ■ Diabetes Mellitus With Hyperglycemia 32 (21) 27 (19) 1.12 (0.63-1.98) ■ Falls 26 (17) 20 (14) 1.24 (0.66-2.33) ▫ Fracture 8 (5) 12 (8) 0.59 (0.23-1.50) ▫ No fracture 18 (12) 8 (6) 2.22 (0.93-5.27) ■ Blood Test Results ▫ Serum urea nitrogen/ creatinine ratio, “20:1 79 (51) 68 (48) 1.15 (0.73-1.81) ▫ Electrolyte panel (imbalance) 51 (33) 58 (41) 0.72 (0.45-1.15) ▫ Serum albumin, <3.0 g/dL 54 (35) 37 (26) 1.53 (0.93-2.53) * Data are presented as No. (%) unless indicated otherwise. † Statistically significant difference (P<.05). ‡ Bladder catheterization was provided in the emergency department or on hospital admission; antibiotic treatment was dispensed in the emergency department. Subjects in the nonbacteriuria control group received antibiotic therapy for other infections. § Information on ischuria was available for 139 of 154 subjects in the bacteriuria group, but it was available for all 142 subjects in the nonbacteriuria control group . ▫ received antibiotic treatment levofloxacin in the ED (OR, 2.09; 95% CI, 1.10-3.96) Although the prevalence of cognitive impairment ranged from 27% to 31% in both study groups, patients with bac- teriuria had a higher rate of delirium on admission than nonbacteriuria controls. The bacteriuria group had a higher, but not statistically significant, frequency of falls without fracture compared to the control group (OR, 2.22; 95% CI, 0.93-5.27). Ninety-two (60%) of the 154 bacteriuria subjects received antibiotic treatment in the ED, with the fluoroquinolone xxx- Xxx et al • Original Contribution ofloxacin as the medication most frequently dispensed. There was a similar frequency of pulmonary disease and diabetes mellitus with hyperglycemia as well as similar results for serum urea nitrogen / creatinine ratio and elec- trolyte panel between the two study groups in the acute discharge diagnosis. However, bacteriuria subjects had a lower frequency of acute pulmonary disease than subjects in the control group (OR, 0.53; 95% CI, 0.32-0.86). ■ Patients with UTI and ASB—By way of comparison, the group of 28 individuals whose UTI diagnosis from exam- iner 2 was under dispute were reassigned to the ASB group because they had symptoms that could overlap with and pos- sibly be explained by acute coexisting illness. As shown in Table 2, patients with UTI and ASB had similar clinical profiles with respect to mean (SD) age, sex, nursing home residency, and medical history. On acute discharge diagnosis, delirium was more common in UTI vs ASB subjects. In addition, UTI subjects had a lower rate of pulmonary disease than those with ASB (OR, 0.14; 95% CI, 0.07-0.31). A similar finding was observed with regard to a diagnosis of pneumonia. In addition, UTI subjects had shorter hospital stays than those with ASB. However, this finding did not remain sta- tistically significant after adjusting for delirium, falls regard- less of fracture status, and pulmonary disease (OR, 0.88; 95% CI, 0.768-1.003). In distinguishing UTI from ASB, pyuria with a urine leukocyte count of at least 2 neutrophils per high power field of spin urine was found in 95 UTI subjects (91%) compared to 34 ASB subjects (68%) (OR, 5.59; 95% CI, 2.19-14.23) with a sensitivity of 91% and a specificity of 32%. The positive and negative predictive values of pyuria for UTI were 74% (95 of 129) and 64% (16 of 25), respectively. Our study revealed consistent interexpert diagnostic agree- ment on UTI. However, the distinction between UTI and ASB was problematic in some cases where acute comorbidity was present (eg, acute pulmonary disease without local urinary tract symptoms), probably as a result of symptom crossover. Pyuria was a sensitive, but not a specific, finding for diag- nosing UTI, which suggests that urinalysis is a good tool for screening patients when there is a high level of suspicion for UTI. In our study, women who were nursing home residents had a higher frequency of bacteriuria—but not UTI—than control group subjects. Delirium was more commonly found in bacteriuria patients than in nonbacteriuria control group subjects. Although delirium and falls without fracture were more common in UTI patients than in those with ASB, it should be noted that no association was observed between delirium and Xxx et al • Original Contribution falls in the bacteriuria vs control comparison regardless of fracture status (OR, 1.12; 95% CI, 0.51-2.49). Our study demonstrated how a potential comorbidity can blur the distinction between UTI and ASB when physicians attempt to diagnose patients who do not present with local uri- nary tract symptoms. The problem of distinguishing UTI from ASB shown in the present study could be the result of several methodologic challenges. For example, the broad definition of UTI we adopted and the potential for crossover symptoms from other acute illnesses in this patient population (eg, nausea and emesis in pulmonary disease) may have influenced the outcomes of the present study. Because of the relatively small sample size and lack of racial and ethnic diversity in our sample population, this diag- nostic challenge requires further investigation—preferably in the form of a randomized double-blind clinical trial of antibi- otic treatment for patients with atypical symptoms and bac- teriuria.16 In the present study, individuals in the nonbacteriuria control group had a higher rate of acute pulmonary disease when compared to subjects in the bacteriuria group. And yet, the coexistence of bacteriuria and acute pulmonary disease could also reflect the high prevalence of ASB in the elderly pop- ulation, especially among functionally impaired nursing home residents.2,17-18 Study wide, among the 101 inpatients with pulmonary disease in the acute discharge diagnosis, 42 had bac- teriuria—a result that is similar to previous reports. 2 A previous surveillance study 19 revealed a 70% preva- lence rate among patients with bacteriuria for no urogenital tract symptoms. This finding was similar to a previous study we conducted in which 70% of hospitalized older adults with bacteriuria had no local urinary tract symptoms or fever (X-X. Xxx, MD, PhD, unpublished data, June 2007). One possible explanation for this phenomenon is that patients with cognitive impairment, a condition that increases in prevalence with age, could be difficult to diagnose because of hampered symptom identification. Clinicians rely on clin- ical symptoms and signs—such as lower abdominal discom- fort or pain, nausea, emesis, fever, and mental status change— when considering diagnostic and treatment options. All too often, however, the symptoms noted by patients in this demo- graphic group are nonspecific and overlap other acute comor- bidities. An algorithmic approach and clinical protocol has been proposed and tested for the management of UTIs in long- term care residents. 20-21 It is not known whether this algo- rithm can be applied to hospitalized older adults. Because of the unique challenges involved in distin- guishing UTI from ASB in older adults,22 a proposal to use “a different combination of existing clinical criteria and geriatric manifestations” to define UTI in older adults has been sug- gested.14 The particular challenges of differentiating the clin- ical manifestations of UTI from potential comorbidities in this JAOA • Vol 109 • No 4 • April 2009 • 223 ■ Mean (SD) Age, y 83 (8) 83 (8) ... ■ Female 86 (83) 44 (88) 0.65 (0.24-1.76) ■ Nursing Home Resident 42 (40) 26 (52) 0.62 (0.31-1.25) ■ Medical History ▫ Cognitive impairment 31 (30) 16 (32) 0.90 (0.44-1.87) ▫ Cardiovascular health – Congestive heart failure 30 (29) 19 (38) 0.66 (0.32-1.35) – Hypertension 65 (63) 28 (56) 1.31 (0.66-2.60) – Stroke 22 (21) 8 (16) 1.41 (0.58-3.43) ▫ Diabetes mellitus 26 (25) 19 (38) 0.54 (0.26-1.27) ▫ Falls 14 (13) 2 (4) 3.73 (0.81-17.11) ■ Hospital Care† ▫ Bladder catheter 55 (53) 23 (46) 1.32 (0.67-2.59) ▫ Antibiotic treatment 67 (64) 25 (50) 1.81 (0.91-3.59) ▫ Mean (SD) length of hospital stay, d 5.0 (2.4) 6.3 (3.4) 0.85 (0.75-0.96)‡ ■ Pyuria§ 95 (91) 34 (68) 5.59 (2.19-14.23)‡ ■ Delirium 40 (38) 6 (12) 4.58 (1.79-11.73)‡ ■ Heart Disease 32 (31) 21 (42) 0.61 (0.31-1.23) ■ Pulmonary Disease 15 (14) 27 (54) 0.14 (0.07-0.31)‡ ▫ Pneumonia 5 (5) 18 (36) 0.09 (0.03-0.26) ‡ ■ Diabetes Mellitus With Hyperglycemia 19 (18) 13 (26) 0.64 (0.28-1.42) ■ Anemia 29 (28) 9 (18) 1.76 (0.76-4.08) ■ Falls 24 (23) 2 (4) 7.20 (1.63-31.83)‡ ▫ Fracture 7 (7) 1 (2) 3.54 (0.42-29.56) ▫ No fracture 17 (16) 1 (2) 9.57 (1.24-74.15)‡ ■ Blood Test Results ▫ Serum urea nitrogen/ creatinine ratio, “20:1 50 (48) 29 (58) 0.67 (0.34-1.32) ▫ Electrolyte panel (imbalance) 32 (31) 19 (38) 0.73 (0.36-1.47) ▫ Serum albumin, <3.0 g/dL 36 (35) 18 (36) 0.94 (0.47-1.90) ■ Urinalysis Results// ▫ Leukocyte esterase (positive) 76 (74) 33 (66) 1.45 (0.70-3.01) ▫ Nitrite (positive) 55 (53) 17 (50) 1.15 (0.53-2.49) * Data are presented as No. (%) unless indicated otherwise. † Bladder catheterization was provided in the emergency department or on hospital admission; antibiotic treatment was dispensed in the emergency department. ‡ Statistically significant difference (P<.05). § Diagnosis of pyuria was based on microscopy (ie, urine leukocyte count “2 neutrophils per high power field of spun urine). // Complete laboratory studies were not available for all study subjects. In the urinary tract infection group, leukocyte esterase dipstick test results were available for 103 of the 104 subjects. Although these results were available for all 50 asymptomatic bacteriuria subjects, nitrite results were available for only 34 of these individuals. population, especially in the context of hospital and long-term care settings, will need to be addressed in any patient care recommendations. Researchers have shown that UTI is a risk factor for the diagnosis of delirium among older adults who are either hos- pitalized23 or receiving care in a psychogeriatric unit. 24 Our findings suggest that older patients exhibiting delirium or a change in mental status receive urinalysis and urine culture to exclude UTI in differential diagnoses. 25 As noted elsewhere, data from our study reveal delirium and greater incidence of falls among patients with UTI. The increased frequency of delirium was seen in the bacteriuria Xxx et al • Original Contribution group when compared to the nonbacteriuria control group, suggesting that delirium was associated with bacteriuria rather than a selection bias. Although a higher frequency of falls without fracture was seen in the discharge diagnosis for the bacteriuria group, the 95% CI ranged between 0.93 and 5.27, which could be a function of the small sample size and lack of power in our study. Previous studies 26-27 have suggested a possible indirect link between UTI and falls. To our awareness, no previous studies have demonstrated a direct association. Further studies using National Hospital Discharge Survey Data or a prospective follow-up study may help establish an association. Evidence is mounting that the increased use of fluoro- quinolone antibiotics is associated with the development of Clostridium difficile colitis.28-29 Therefore, it is important that physicians take care to distinguish UTI from ASB. Current clinical guidelines2 recommend against the use of antibiotic treatment for ASB. In addition to the sample size and demographic limitations previously noted, when evaluating the general applicability of our results, readers may wish to consider certain method- ologic limitations of our study. First, the “other symptoms” used for the definition of UTI10—beyond local symptoms—are not yet well established in the medical literature. Second, it remains to be verified whether patients with bacteriuria and falls at admission diagnosis should be automatically diag- nosed with UTI. This conclusion is not supported by our study’s findings. In designing our study, we included falls as a manifestation of UTI mainly because this criterion was used by previous researchers.10,12 Also, data from urinalysis and urine cultures reviewed in this study were not obtained during the course of routine surveillance procedures. As a result, it is possible that our sample population is subject to selection bias. On a related note, our study design did not afford us with the opportu- nity to repeat urinalysis and urine culture to confirm ASB diagnosis as recommended by current clinical practice guide- lines.2 Finally, we cannot rule out a type I error in the association between bacteriuria status and acute pulmonary disease among our subjects.

Patient characteristics. Two comparisons were made for the purposes of this inves- tigation: (1) patients with bacteriuria vs nonbacteriuria control subjects (Table 1), and (2) UTI vs ASB subjects (Table 2). All 296 study subjects were white. ■ Patients with bacteriuria vs control subjects —When com- pared with individuals from the nonbacteriuria control group ( (Table 1), patients with bacteriuria (either UTI or ASB) were more often: ▫ female (84% vs 75%; OR, 1.84; 95% CI, 1.03-3.26) ▫ nursing home residents (OR, 1.74; 95% CI, 1.08-2.80) ▫ had a history of hypertension (OR, 1.81; 95% CI, 1.14-2.87) ▫ received bladder catheterization in the ED or on admission (OR, 1.72; 95% CI, 1.08-2.74) JAOA • Vol 109 • No 4 • April 2009 • 221 ■ Mean (SD) Age, y 83 (8) 82 (8) ... ■ Female 130 (84) 106 (75) 1.84 (1.03-3.26)† ■ Nursing Home Resident 68 (44) 47 (33) 1.74 (1.08-2.80)† ■ Medical History ▫ Cognitive impairment 47 (31) 38 (27) 1.20 (0.73-1.99) ▫ Cardiovascular health – Hypertension 93 (60) 65 (46) 1.81 (1.14-2.87)† – Stroke 30 (19) 34 (24) 0.77 (0.44-1.34) ▫ Pulmonary disease – Chronic obstructive pulmonary disease 33 (21) 36 (25) 0.80 (0.47-1.38) – Pneumonia 16 (10) 19 (13) 0.75 (0.37-1.52) ▫ Diabetes mellitus 45 (29) 34 (24) 1.31 (0.78-2.20) ▫ Falls 16 (10) 9 (6) 1.71 (0.73-4.01) ■ Hospital Care‡ ▫ Bladder catheter 78 (51) 53 (37) 1.72 (1.08-2.74)† ▫ Antibiotic treatment 92 (60) 67 (47) 1.66 (1.05-2.63) † – Levofloxacin treament 53 (58) 26 (39) 2.09 (1.10-3.96)† ■ Ischuria§ 5 (4) 11 (8) 0.44 (0.15-1.31) ■ Delirium 46 (30) 11 (8) 5.07 (2.51-10.27)† ■ Cardiovascular Health ▫ Atrial fibrillation 19 (12) 19 (13) 0.91 (0.46-1.80) ▫ Congestive heart failure 31 (20) 18 (13) 1.74 (0.92-3.27) ■ Pulmonary Disease 42 (27) 59 (42) 0.53 (0.32-0.86)† ▫ Pneumonia 23 (15) 32 (23) 0.60 (0.33-1.09) ■ Diabetes Mellitus With Hyperglycemia 32 (21) 27 (19) 1.12 (0.63-1.98) ■ Falls 26 (17) 20 (14) 1.24 (0.66-2.33) ▫ Fracture 8 (5) 12 (8) 0.59 (0.23-1.50) ▫ No fracture 18 (12) 8 (6) 2.22 (0.93-5.27) ■ Blood Test Results ▫ Serum urea nitrogen/ creatinine ratio, “20:1 79 (51) 68 (48) 1.15 (0.73-1.81) ▫ Electrolyte panel (imbalance) 51 (33) 58 (41) 0.72 (0.45-1.15) ▫ Serum albumin, <3.0 g/dL 54 (35) 37 (26) 1.53 (0.93-2.53) * Data are presented as No. (%) unless indicated otherwise. † Statistically significant difference (( P<.05). ‡ Bladder catheterization was provided in the emergency department or on hospital admission; antibiotic treatment was dispensed in the emergency department. Subjects in the nonbacteriuria control group received antibiotic therapy for other infections. § Information on ischuria was available for 139 of 154 subjects in the bacteriuria group, but it was available for all 142 subjects in the nonbacteriuria control group . ▫ received antibiotic treatment levofloxacin in the ED (OR, 2.09; 95% CI, 1.10-3.96) Although the prevalence of cognitive impairment ranged from 27% to 31% in both study groups, patients with bac- teriuria had a higher rate of delirium on admission than nonbacteriuria controls. The bacteriuria group had a higher, but not statistically significant, frequency of falls without fracture compared to the control group (OR, 2.22; 95% CI, 0.93-5.27). Ninety-two (60%) of the 154 bacteriuria subjects received antibiotic treatment in the ED, with the fluoroquinolone xxx- Xxx et al • Original Contribution lev- ofloxacin as the medication most frequently dispensed. There was a similar frequency of pulmonary disease and diabetes mellitus with hyperglycemia as well as similar results for serum urea nitrogen / creatinine ratio and elec- trolyte panel between the two study groups in the acute discharge diagnosis. However, bacteriuria subjects had a lower frequency of acute pulmonary disease than subjects in the control group (OR, 0.53; 95% CI, 0.32-0.86). ■ Patients with UTI and ASB—By way of comparison, the group of 28 individuals whose UTI diagnosis from exam- iner 2 was under dispute were reassigned to the ASB group because they had symptoms that could overlap with and pos- sibly be explained by acute coexisting illness. As shown in Table 2, patients with UTI and ASB had similar clinical profiles with respect to mean (SD) age, sex, nursing home residency, and medical history. On acute discharge diagnosis, delirium was more common in UTI vs ASB subjects. In addition, UTI subjects had a lower rate of pulmonary disease than those with ASB (OR, 0.14; 95% CI, 0.07-0.31). A similar finding was observed with regard to a diagnosis of pneumonia. In addition, UTI subjects had shorter hospital stays than those with ASB. However, this finding did not remain sta- tistically significant after adjusting for delirium, falls regard- less of fracture status, and pulmonary disease (OR, 0.88; 95% CI, 0.768-1.003). In distinguishing UTI from ASB, pyuria with a urine leukocyte count of at least 2 neutrophils per high power field of spin urine was found in 95 UTI subjects (91%) compared to 34 ASB subjects (68%) (OR, 5.59; 95% CI, 2.19-14.23) with a sensitivity of 91% and a specificity of 32%. The positive and negative predictive values of pyuria for UTI were 74% (95 of 129) and 64% (16 of 25), respectively. Our study revealed consistent interexpert diagnostic agree- ment on UTI. However, the distinction between UTI and ASB was problematic in some cases where acute comorbidity was present (eg, acute pulmonary disease without local urinary tract symptoms), probably as a result of symptom crossover. Pyuria was a sensitive, but not a specific, finding for diag- nosing UTI, which suggests that urinalysis is a good tool for screening patients when there is a high level of suspicion for UTI. In our study, women who were nursing home residents had a higher frequency of bacteriuria—but not UTI—than control group subjects. Delirium was more commonly found in bacteriuria patients than in nonbacteriuria control group subjects. Although delirium and falls without fracture were more common in UTI patients than in those with ASB, it should be noted that no association was observed between delirium and Xxx et al • Original Contribution falls in the bacteriuria vs control comparison regardless of fracture status (OR, 1.12; 95% CI, 0.51-2.49). Our study demonstrated how a potential comorbidity can blur the distinction between UTI and ASB when physicians attempt to diagnose patients who do not present with local uri- nary tract symptoms. The problem of distinguishing UTI from ASB shown in the present study could be the result of several methodologic challenges. For example, the broad definition of UTI we adopted and the potential for crossover symptoms from other acute illnesses in this patient population (eg, nausea and emesis in pulmonary disease) may have influenced the outcomes of the present study. Because of the relatively small sample size and lack of racial and ethnic diversity in our sample population, this diag- nostic challenge requires further investigation—preferably in the form of a randomized double-blind clinical trial of antibi- otic treatment for patients with atypical symptoms and bac- teriuria.16 In the present study, individuals in the nonbacteriuria control group had a higher rate of acute pulmonary disease when compared to subjects in the bacteriuria group. And yet, the coexistence of bacteriuria and acute pulmonary disease could also reflect the high prevalence of ASB in the elderly pop- ulation, especially among functionally impaired nursing home residents.2,17-18 Study wide, among the 101 inpatients with pulmonary disease in the acute discharge diagnosis, 42 had bac- teriuria—a result that is similar to previous reports. 2 reports.2 A previous surveillance study 19 study19 revealed a 70% preva- lence rate among patients with bacteriuria for no urogenital tract symptoms. This finding was similar to a previous study we conducted in which 70% of hospitalized older adults with bacteriuria had no local urinary tract symptoms or fever (X-X. Xxx, MD, PhD, unpublished data, June 2007). One possible explanation for this phenomenon is that patients with cognitive impairment, a condition that increases in prevalence with age, could be difficult to diagnose because of hampered symptom identification. Clinicians rely on clin- ical symptoms and signs—such as lower abdominal discom- fort or pain, nausea, emesis, fever, and mental status change— when considering diagnostic and treatment options. All too often, however, the symptoms noted by patients in this demo- graphic group are nonspecific and overlap other acute comor- bidities. An algorithmic approach and clinical protocol has been proposed and tested for the management of UTIs in long- term care residents. 20residents.20-21 It is not known whether this algo- rithm can be applied to hospitalized older adults. Because of the unique challenges involved in distin- guishing UTI from ASB in older adults,22 a proposal to use “a different combination of existing clinical criteria and geriatric manifestations” to define UTI in older adults has been sug- gested.14 The particular challenges of differentiating the clin- ical manifestations of UTI from potential comorbidities in this JAOA • Vol 109 • No 4 • April 2009 • 223 ■ Mean (SD) Age, y 83 (8) 83 (8) ... ■ Female 86 (83) 44 (88) 0.65 (0.24-1.76) ■ Nursing Home Resident 42 (40) 26 (52) 0.62 (0.31-1.25) ■ Medical History ▫ Cognitive impairment 31 (30) 16 (32) 0.90 (0.44-1.87) ▫ Cardiovascular health – Congestive heart failure 30 (29) 19 (38) 0.66 (0.32-1.35) – Hypertension 65 (63) 28 (56) 1.31 (0.66-2.60) – Stroke 22 (21) 8 (16) 1.41 (0.58-3.43) ▫ Diabetes mellitus 26 (25) 19 (38) 0.54 (0.26-1.27) ▫ Falls 14 (13) 2 (4) 3.73 (0.81-17.11) ■ Hospital Care† ▫ Bladder catheter 55 (53) 23 (46) 1.32 (0.67-2.59) ▫ Antibiotic treatment 67 (64) 25 (50) 1.81 (0.91-3.59) ▫ Mean (SD) length of hospital stay, d 5.0 (2.4) 6.3 (3.4) 0.85 (0.75-0.96)‡ ■ Pyuria§ 95 (91) 34 (68) 5.59 (2.19-14.23)‡ ■ Delirium 40 (38) 6 (12) 4.58 (1.79-11.73)‡ ■ Heart Disease 32 (31) 21 (42) 0.61 (0.31-1.23) ■ Pulmonary Disease 15 (14) 27 (54) 0.14 (0.07-0.31)‡ ▫ Pneumonia 5 (5) 18 (36) 0.09 (0.03-0.26) ‡ ■ Diabetes Mellitus With Hyperglycemia 19 (18) 13 (26) 0.64 (0.28-1.42) ■ Anemia 29 (28) 9 (18) 1.76 (0.76-4.08) ■ Falls 24 (23) 2 (4) 7.20 (1.63-31.83)‡ ▫ Fracture 7 (7) 1 (2) 3.54 (0.42-29.56) ▫ No fracture 17 (16) 1 (2) 9.57 (1.24-74.15)‡ ■ Blood Test Results ▫ Serum urea nitrogen/ creatinine ratio, “20:1 50 (48) 29 (58) 0.67 (0.34-1.32) ▫ Electrolyte panel (imbalance) 32 (31) 19 (38) 0.73 (0.36-1.47) ▫ Serum albumin, <3.0 g/dL 36 (35) 18 (36) 0.94 (0.47-1.90) ■ Urinalysis Results// ▫ Leukocyte esterase (positive) 76 (74) 33 (66) 1.45 (0.70-3.01) ▫ Nitrite (positive) 55 (53) 17 (50) 1.15 (0.53-2.49) * Data are presented as No. (%) unless indicated otherwise. † Bladder catheterization was provided in the emergency department or on hospital admission; antibiotic treatment was dispensed in the emergency department. ‡ Statistically significant difference (( P<.05). § Diagnosis of pyuria was based on microscopy (ie, urine leukocyte count “2 neutrophils per high power field of spun urine). // Complete laboratory studies were not available for all study subjects. In the urinary tract infection group, leukocyte esterase dipstick test results were available for 103 of the 104 subjects. Although these results were available for all 50 asymptomatic bacteriuria subjects, nitrite results were available for only 34 of these individuals. population, especially in the context of hospital and long-term care settings, will need to be addressed in any patient care recommendations. Researchers have shown that UTI is a risk factor for the diagnosis of delirium among older adults who are either hos- pitalized23 or receiving care in a psychogeriatric unit. 24 unit.24 Our findings suggest that older patients exhibiting delirium or a change in mental status receive urinalysis and urine culture to exclude UTI in differential diagnoses. 25 diagnoses.25 As noted elsewhere, data from our study reveal delirium and greater incidence of falls among patients with UTI. The increased frequency of delirium was seen in the bacteriuria Xxx et al • Original Contribution group when compared to the nonbacteriuria control group, suggesting that delirium was associated with bacteriuria rather than a selection bias. Although a higher frequency of falls without fracture was seen in the discharge diagnosis for the bacteriuria group, the 95% CI ranged between 0.93 and 5.27, which could be a function of the small sample size and lack of power in our study. Previous studies 26studies26-27 have suggested a possible indirect link between UTI and falls. To our awareness, no previous studies have demonstrated a direct association. Further studies using National Hospital Discharge Survey Data or a prospective follow-up study may help establish an association. Evidence is mounting that the increased use of fluoro- quinolone antibiotics is associated with the development of Clostridium difficile colitis.28-29 Therefore, it is important that physicians take care to distinguish UTI from ASB. Current clinical guidelines2 recommend against the use of antibiotic treatment for ASB. In addition to the sample size and demographic limitations previously noted, when evaluating the general applicability of our results, readers may wish to consider certain method- ologic limitations of our study. First, the “other symptoms” used for the definition of UTI10—beyond local symptoms—are not yet well established in the medical literature. Second, it remains to be verified whether patients with bacteriuria and falls at admission diagnosis should be automatically diag- nosed with UTI. This conclusion is not supported by our study’s findings. In designing our study, we included falls as a manifestation of UTI mainly because this criterion was used by previous researchers.10,12 Also, data from urinalysis and urine cultures reviewed in this study were not obtained during the course of routine surveillance procedures. As a result, it is possible that our sample population is subject to selection bias. On a related note, our study design did not afford us with the opportu- nity to repeat urinalysis and urine culture to confirm ASB diagnosis as recommended by current clinical practice guide- lines.2 Finally, we cannot rule out a type I error in the association between bacteriuria status and acute pulmonary disease among our subjects.

Patient characteristics. Two comparisons were made for the purposes of this inves- tigation: (1) patients with bacteriuria vs nonbacteriuria control subjects (Table 1), and (2) UTI vs ASB subjects (Table 2). All 296 study subjects were white. ■ Patients with bacteriuria vs control subjects —When com- pared with individuals from the nonbacteriuria control group ( Table 1), patients with bacteriuria (either UTI or ASB) were more often: ▫ female (84% vs 75%; OR, 1.84; 95% CI, 1.03-3.26) ▫ nursing home residents (OR, 1.74; 95% CI, 1.08-2.80) ▫ had a history of hypertension (OR, 1.81; 95% CI, 1.14-2.87) ▫ received bladder catheterization in the ED or on admission (OR, 1.72; 95% CI, 1.08-2.74) JAOA • Vol 109 • No 4 • April 2009 • 221 ■ Mean (SD) Age, y 83 (8) 82 (8) ... ■ Female 130 (84) 106 (75) 1.84 (1.03-3.26)† ■ Nursing Home Resident 68 (44) 47 (33) 1.74 (1.08-2.80)† ■ Medical History ▫ Cognitive impairment 47 (31) 38 (27) 1.20 (0.73-1.99) ▫ Cardiovascular health – Hypertension 93 (60) 65 (46) 1.81 (1.14-2.87)† – Stroke 30 (19) 34 (24) 0.77 (0.44-1.34) ▫ Pulmonary disease – Chronic obstructive pulmonary disease 33 (21) 36 (25) 0.80 (0.47-1.38) – Pneumonia 16 (10) 19 (13) 0.75 (0.37-1.52) ▫ Diabetes mellitus 45 (29) 34 (24) 1.31 (0.78-2.20) ▫ Falls 16 (10) 9 (6) 1.71 (0.73-4.01) ■ Hospital Care‡ ▫ Bladder catheter 78 (51) 53 (37) 1.72 (1.08-2.74)† ▫ Antibiotic treatment 92 (60) 67 (47) 1.66 (1.05-2.63) † – Levofloxacin treament 53 (58) 26 (39) 2.09 (1.10-3.96)† ■ Ischuria§ 5 (4) 11 (8) 0.44 (0.15-1.31) ■ Delirium 46 (30) 11 (8) 5.07 (2.51-10.27)† ■ Cardiovascular Health ▫ Atrial fibrillation 19 (12) 19 (13) 0.91 (0.46-1.80) ▫ Congestive heart failure 31 (20) 18 (13) 1.74 (0.92-3.27) ■ Pulmonary Disease 42 (27) 59 (42) 0.53 (0.32-0.86)† ▫ Pneumonia 23 (15) 32 (23) 0.60 (0.33-1.09) ■ Diabetes Mellitus With Hyperglycemia 32 (21) 27 (19) 1.12 (0.63-1.98) ■ Falls 26 (17) 20 (14) 1.24 (0.66-2.33) ▫ Fracture 8 (5) 12 (8) 0.59 (0.23-1.50) ▫ No fracture 18 (12) 8 (6) 2.22 (0.93-5.27) ■ Blood Test Results ▫ Serum urea nitrogen/ creatinine ratio, “20:1 79 (51) 68 (48) 1.15 (0.73-1.81) ▫ Electrolyte panel (imbalance) 51 (33) 58 (41) 0.72 (0.45-1.15) ▫ Serum albumin, <3.0 g/dL 54 (35) 37 (26) 1.53 (0.93-2.53) * Data are presented as No. (%) unless indicated otherwise. † Statistically significant difference (P<.05). ‡ Bladder catheterization was provided in the emergency department or on hospital admission; antibiotic treatment was dispensed in the emergency department. Subjects in the nonbacteriuria control group received antibiotic therapy for other infections. § Information on ischuria was available for 139 of 154 subjects in the bacteriuria group, but it was available for all 142 subjects in the nonbacteriuria control group . ▫ received antibiotic treatment levofloxacin in the ED (OR, 2.09; 95% CI, 1.10-3.96) Although the prevalence of cognitive impairment ranged from 27% to 31% in both study groups, patients with bac- teriuria had a higher rate of delirium on admission than nonbacteriuria controls. The bacteriuria group had a higher, but not statistically significant, frequency of falls without fracture compared to the control group (OR, 2.22; 95% CI, 0.93-5.27). Ninety-two (60%) of the 154 bacteriuria subjects received antibiotic treatment in the ED, with the fluoroquinolone xxx- Xxx et al • Original Contribution lev- ofloxacin as the medication most frequently dispensed. There was a similar frequency of pulmonary disease and diabetes mellitus with hyperglycemia as well as similar results for serum urea nitrogen / creatinine ratio and elec- trolyte panel between the two study groups in the acute discharge diagnosis. However, bacteriuria subjects had a lower frequency of acute pulmonary disease than subjects in the control group (OR, 0.53; 95% CI, 0.32-0.86). ■ Patients with UTI and ASB—By way of comparison, the group of 28 individuals whose UTI diagnosis from exam- iner 2 was under dispute were reassigned to the ASB group because they had symptoms that could overlap with and pos- sibly be explained by acute coexisting illness. As shown in Table 2, patients with UTI and ASB had similar clinical profiles with respect to mean (SD) age, sex, nursing home residency, and medical history. On acute discharge diagnosis, delirium was more common in UTI vs ASB subjects. In addition, UTI subjects had a lower rate of pulmonary disease than those with ASB (OR, 0.14; 95% CI, 0.07-0.31). A similar finding was observed with regard to a diagnosis of pneumonia. In addition, UTI subjects had shorter hospital stays than those with ASB. However, this finding did not remain sta- tistically significant after adjusting for delirium, falls regard- less of fracture status, and pulmonary disease (OR, 0.88; 95% CI, 0.768-1.003). In distinguishing UTI from ASB, pyuria with a urine leukocyte count of at least 2 neutrophils per high power field of spin urine was found in 95 UTI subjects (91%) compared to 34 ASB subjects (68%) (OR, 5.59; 95% CI, 2.19-14.23) with a sensitivity of 91% and a specificity of 32%. The positive and negative predictive values of pyuria for UTI were 74% (95 of 129) and 64% (16 of 25), respectively. Our study revealed consistent interexpert diagnostic agree- ment on UTI. However, the distinction between UTI and ASB was problematic in some cases where acute comorbidity was present (eg, acute pulmonary disease without local urinary tract symptoms), probably as a result of symptom crossover. Pyuria was a sensitive, but not a specific, finding for diag- nosing UTI, which suggests that urinalysis is a good tool for screening patients when there is a high level of suspicion for UTI. In our study, women who were nursing home residents had a higher frequency of bacteriuria—but not UTI—than control group subjects. Delirium was more commonly found in bacteriuria patients than in nonbacteriuria control group subjects. Although delirium and falls without fracture were more common in UTI patients than in those with ASB, it should be noted that no association was observed between delirium and Xxx et al • Original Contribution falls in the bacteriuria vs control comparison regardless of fracture status (OR, 1.12; 95% CI, 0.51-2.49). Our study demonstrated how a potential comorbidity can blur the distinction between UTI and ASB when physicians attempt to diagnose patients who do not present with local uri- nary tract symptoms. The problem of distinguishing UTI from ASB shown in the present study could be the result of several methodologic challenges. For example, the broad definition of UTI we adopted and the potential for crossover symptoms from other acute illnesses in this patient population (eg, nausea and emesis in pulmonary disease) may have influenced the outcomes of the present study. Because of the relatively small sample size and lack of racial and ethnic diversity in our sample population, this diag- nostic challenge requires further investigation—preferably in the form of a randomized double-blind clinical trial of antibi- otic treatment for patients with atypical symptoms and bac- teriuria.16 In the present study, individuals in the nonbacteriuria control group had a higher rate of acute pulmonary disease when compared to subjects in the bacteriuria group. And yet, the coexistence of bacteriuria and acute pulmonary disease could also reflect the high prevalence of ASB in the elderly pop- ulation, especially among functionally impaired nursing home residents.2,17-18 Study wide, among the 101 inpatients with pulmonary disease in the acute discharge diagnosis, 42 had bac- teriuria—a result that is similar to previous reports. 2 A previous surveillance study 19 revealed a 70% preva- lence rate among patients with bacteriuria for no urogenital tract symptoms. This finding was similar to a previous study we conducted in which 70% of hospitalized older adults with bacteriuria had no local urinary tract symptoms or fever (X-X. Xxx, MD, PhD, unpublished data, June 2007). One possible explanation for this phenomenon is that patients with cognitive impairment, a condition that increases in prevalence with age, could be difficult to diagnose because of hampered symptom identification. Clinicians rely on clin- ical symptoms and signs—such as lower abdominal discom- fort or pain, nausea, emesis, fever, and mental status change— when considering diagnostic and treatment options. All too often, however, the symptoms noted by patients in this demo- graphic group are nonspecific and overlap other acute comor- bidities. An algorithmic approach and clinical protocol has been proposed and tested for the management of UTIs in long- term care residents. 20-21 It is not known whether this algo- rithm can be applied to hospitalized older adults. Because of the unique challenges involved in distin- guishing UTI from ASB in older adults,22 a proposal to use “a different combination of existing clinical criteria and geriatric manifestations” to define UTI in older adults has been sug- gested.14 The particular challenges of differentiating the clin- ical manifestations of UTI from potential comorbidities in this JAOA • Vol 109 • No 4 • April 2009 • 223 ■ Mean (SD) Age, y 83 (8) 83 (8) ... ■ Female 86 (83) 44 (88) 0.65 (0.24-1.76) ■ Nursing Home Resident 42 (40) 26 (52) 0.62 (0.31-1.25) ■ Medical History ▫ Cognitive impairment 31 (30) 16 (32) 0.90 (0.44-1.87) ▫ Cardiovascular health – Congestive heart failure 30 (29) 19 (38) 0.66 (0.32-1.35) – Hypertension 65 (63) 28 (56) 1.31 (0.66-2.60) – Stroke 22 (21) 8 (16) 1.41 (0.58-3.43) ▫ Diabetes mellitus 26 (25) 19 (38) 0.54 (0.26-1.27) ▫ Falls 14 (13) 2 (4) 3.73 (0.81-17.11) ■ Hospital Care† ▫ Bladder catheter 55 (53) 23 (46) 1.32 (0.67-2.59) ▫ Antibiotic treatment 67 (64) 25 (50) 1.81 (0.91-3.59) ▫ Mean (SD) length of hospital stay, d 5.0 (2.4) 6.3 (3.4) 0.85 (0.75-0.96)‡ ■ Pyuria§ 95 (91) 34 (68) 5.59 (2.19-14.23)‡ ■ Delirium 40 (38) 6 (12) 4.58 (1.79-11.73)‡ ■ Heart Disease 32 (31) 21 (42) 0.61 (0.31-1.23) ■ Pulmonary Disease 15 (14) 27 (54) 0.14 (0.07-0.31)‡ ▫ Pneumonia 5 (5) 18 (36) 0.09 (0.03-0.26) ‡ ■ Diabetes Mellitus With Hyperglycemia 19 (18) 13 (26) 0.64 (0.28-1.42) ■ Anemia 29 (28) 9 (18) 1.76 (0.76-4.08) ■ Falls 24 (23) 2 (4) 7.20 (1.63-31.83)‡ ▫ Fracture 7 (7) 1 (2) 3.54 (0.42-29.56) ▫ No fracture 17 (16) 1 (2) 9.57 (1.24-74.15)‡ ■ Blood Test Results ▫ Serum urea nitrogen/ creatinine ratio, “20:1 50 (48) 29 (58) 0.67 (0.34-1.32) ▫ Electrolyte panel (imbalance) 32 (31) 19 (38) 0.73 (0.36-1.47) ▫ Serum albumin, <3.0 g/dL 36 (35) 18 (36) 0.94 (0.47-1.90) ■ Urinalysis Results// ▫ Leukocyte esterase (positive) 76 (74) 33 (66) 1.45 (0.70-3.01) ▫ Nitrite (positive) 55 (53) 17 (50) 1.15 (0.53-2.49) * Data are presented as No. (%) unless indicated otherwise. † Bladder catheterization was provided in the emergency department or on hospital admission; antibiotic treatment was dispensed in the emergency department. ‡ Statistically significant difference (P<.05). § Diagnosis of pyuria was based on microscopy (ie, urine leukocyte count “2 neutrophils per high power field of spun urine). // Complete laboratory studies were not available for all study subjects. In the urinary tract infection group, leukocyte esterase dipstick test results were available for 103 of the 104 subjects. Although these results were available for all 50 asymptomatic bacteriuria subjects, nitrite results were available for only 34 of these individuals. population, especially in the context of hospital and long-term care settings, will need to be addressed in any patient care recommendations. Researchers have shown that UTI is a risk factor for the diagnosis of delirium among older adults who are either hos- pitalized23 or receiving care in a psychogeriatric unit. 24 Our findings suggest that older patients exhibiting delirium or a change in mental status receive urinalysis and urine culture to exclude UTI in differential diagnoses. 25 As noted elsewhere, data from our study reveal delirium and greater incidence of falls among patients with UTI. The increased frequency of delirium was seen in the bacteriuria Xxx et al • Original Contribution group when compared to the nonbacteriuria control group, suggesting that delirium was associated with bacteriuria rather than a selection bias. Although a higher frequency of falls without fracture was seen in the discharge diagnosis for the bacteriuria group, the 95% CI ranged between 0.93 and 5.27, which could be a function of the small sample size and lack of power in our study. Previous studies 26-27 have suggested a possible indirect link between UTI and falls. To our awareness, no previous studies have demonstrated a direct association. Further studies using National Hospital Discharge Survey Data or a prospective follow-up study may help establish an association. Evidence is mounting that the increased use of fluoro- quinolone antibiotics is associated with the development of Clostridium difficile colitis.28-29 Therefore, it is important that physicians take care to distinguish UTI from ASB. Current clinical guidelines2 recommend against the use of antibiotic treatment for ASB. In addition to the sample size and demographic limitations previously noted, when evaluating the general applicability of our results, readers may wish to consider certain method- ologic limitations of our study. First, the “other symptoms” used for the definition of UTI10—beyond local symptoms—are not yet well established in the medical literature. Second, it remains to be verified whether patients with bacteriuria and falls at admission diagnosis should be automatically diag- nosed with UTI. This conclusion is not supported by our study’s findings. In designing our study, we included falls as a manifestation of UTI mainly because this criterion was used by previous researchers.10,12 Also, data from urinalysis and urine cultures reviewed in this study were not obtained during the course of routine surveillance procedures. As a result, it is possible that our sample population is subject to selection bias. On a related note, our study design did not afford us with the opportu- nity to repeat urinalysis and urine culture to confirm ASB diagnosis as recommended by current clinical practice guide- lines.2 Finally, we cannot rule out a type I error in the association between bacteriuria status and acute pulmonary disease among our subjects.