Hospitals typically reduce staffing levels and the availability of diagnostic, laboratory, and treatment services on weekends, and patients admitted on weekends exhibit poorer in‐hospital outcomes for several medical conditions.[1, 2, 3, 4, 5, 6, 7, 8, 9] Whether or not patients discharged on weekends have worse clinical outcomes has been less well studied.[10, 11, 12] Discharge rates on Saturday and Sunday are lower than for the other 5 days of the week,[12] but bed shortages and hospital overcrowding have increased the demand for maximizing 24/7 week‐round discharge efficiency. Given that the number of patients discharged on weekends is likely to continue to increase, it is important to assess the risk of weekend discharge on outcomes monitored as performance indicators by organizations such as the Centers for Medicare and Medicaid Services, the American Medical Association Physicians Consortium for Performance Improvement, the National Quality Forum, and the Joint Commission.

Thus, we designed this study to evaluate baseline characteristics, length of stay (LOS), and postdischarge outcomes for general internal medicine (GIM) patients in teaching hospitals discharged on weekends compared to weekdays. Our objective was to determine whether postdischarge outcomes differed for patients discharged on weekends versus weekdays.

METHODS

RESULTS

Weekday Versus Weekend Discharge

Although patients admitted on weekdays and weekends were very similar (data available upon request), patients discharged on weekends (compared to those discharged on weekdays) were younger, more likely to be discharged home without additional support, and had fewer comorbidities (Table 1, Figure 1). Patients discharged on weekends had shorter lengths of stay than those discharged on weekdays (5.6 days vs 7.9 days, P<0.0001). In adjusted linear regression analyses, this 2.3‐day difference remained statistically significant (adjusted P value <0.0001).

Figure 1

Patients discharged on a weekend exhibited lower unadjusted 30‐day rates of death or readmission than those discharged on a weekday (10.6% vs 13.2%), but these differences disappeared after multivariable adjustment that accounted for differences in risk profile (aOR: 0.94, 95% CI: 0.771.16 (Table 2). Results were similar in sensitivity analyses adjusting for LACE scores without LOS included (aOR: 0.88, 95% CI: 0.711.08) or adjusting for LACE scores using expected LOS rather than actual LOS (aOR: 0.90, 95% CI: 0.731.10). Restricting the analysis to only those patients deemed to be at high risk for events due to LACE scores of 10 or greater confirmed that weekend and weekday discharges had similar outcomes in the first 30 days after discharge (aOR: 1.09, 95% CI: 0.851.41, Table 2). Similar patterns were seen when we included ED visits as part of the composite endpoint (ie, death, unplanned readmission, or unplanned ED visit within 30 days of discharge) (Table 2).

Table 2.Postdischarge Outcomes After a General Internal Medicine Hospitalization in a Teaching Hospital

	Weekend Discharge, n/N (%)	Weekday Discharge, n/N (%)	Unadjusted P Value	aOR* (95% CI)	Adjusted P Value
NOTE: Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; ED, emergency department; LACE, length of hospital stay, acuity of admission, comorbidity burden quantified using the Charlson Comorbidity Index, and emergency department visits in the 6 months prior to admission. *Multivariable models adjust for age, sex, hospital, and LACE score at time of discharge from index hospitalization. Weekday discharge is reference group for odds ratios.
Death/readmission within 30 days
All 7 teaching hospitals, all patients	121/1146 (10.6)	901/6845 (13.2)	0.01	0.94 (0.77‐1.16)	0.58
All 7 teaching hospitals, but only patients with LACE <10	37/647 (5.7)	225/2753 (8.2)	0.04	0.72 (0.50, 1.03)	0.07
All 7 teaching hospitals, but only patients with LACE 10	84/499 (16.8)	676/4092 (16.5)	0.86	1.09 (0.85‐1.41)	0.49
Death/readmission/ED visit within 30 days
All 7 teaching hospitals, all patients	218/1146 (19.0)	1445/6845 (21.1)	0.11	0.98 (0.83‐1.15)	0.79
All 7 teaching hospitals, but only patients with LACE <10	90/647 (13.9)	460/2753 (16.7)	0.08	0.83 (0.64‐1.06)	0.13
All 7 teaching hospitals, but only patients with LACE 10	128/499 (25.7)	985/4092 (24.1)	0.44	1.12 (0.90‐1.39)	0.31
Death within 30 days
All 7 teaching hospitals, all patients	24/1146 (2.1)	215/6845 (3.1)	0.05	0.97 (0.63‐1.51)	0.89
All 7 teaching hospitals, but only patients with LACE <10	4/647 (0.6)	23/2753 (0.8)	0.58	0.89 (0.30, 2.62)	0.83
All 7 teaching hospitals, but only patients with LACE 10	20/499 (4.0)	192/4092 (4.7)	0.49	0.99 (0.61‐1.61)	0.98
Readmission within 30 days
All 7 teaching hospitals, all patients	105/1146 (9.2)	751/6845 (11.0)	0.07	0.94 (0.76‐1.17)	0.59
All 7 teaching hospitals, but only patients with LACE <10	33/647 (5.1)	211/2753 (7.7)	0.02	0.68 (0.46‐0.99)	0.04
All 7 teaching hospitals, but only patients with LACE 10	72/499 (14.4)	540/4092 (13.2)	0.44	1.14 (0.87‐1.49)	0.34
ED visit within 30 days
All 7 teaching hospitals, all patients	182/1146 (15.9)	1118/6845 (16.3)	0.70	1.00 (0.84‐1.19)	0.99
All 7 teaching hospitals, but only patients with LACE <10	83/647 (12.8)	412/2753 (15.0)	0.17	0.84 (0.65, 1.09)	0.20
All 7 teaching hospitals, but only patients with LACE 10	99/499 (19.8)	706/4092 (17.3)	0.15	1.17 (0.92‐1.48)	0.20

DISCUSSION

Our data suggest that patients discharged from the GIM teaching wards we studied on weekends were appropriately triaged, as they did not exhibit a higher risk of adverse events postdischarge. Although patients discharged on weekends tended to be younger and had less comorbidities than those discharged during the week, we adjusted for baseline covariates in analyses, and we did not find an association between weekend discharge and increased postdischarge events even among the subset of patients deemed to be at high risk for postdischarge adverse events (based on high LACE scores). To our knowledge, although we previously examined this issue in patients with a most‐responsible diagnosis of heart failure,[10] examining weekend versus weekday discharges in the full gamut of general medical patients admitted to teaching hospitals has not previously been examined.

In our previous study[10] of over 24,000 heart failure patients discharged over 10 years (up to June 2009, therefore no overlap with any patients in this study), we also found that patients discharged on the weekends were younger, had fewer comorbidities, and shorter lengths of stay. Although postdischarge death/readmission rates were higher for weekend discharged patients in our earlier study (21.1% vs 19.5%, adjusted hazard ratio: 1.15, 95% CI: 1.061.25), it is worth noting that this was almost entirely driven by data from nonteaching hospitals and cardiology wards. Thus, it is important to reiterate that the findings in our current study are for GIM wards in teaching hospitals and may not be generalizable to less‐structured nonteaching settings.

Although we did not study physician decision making, our results suggest that physicians are incorporating discharge day into their discharge decision making. They may be selecting younger patients with less comorbidities for weekend discharges, or they may be delaying the discharges of older patients with more comorbidities for weekday discharges. Either is not surprising given the realities of weekend inpatient care: reduced staffing and frequent cross‐coverage (of physicians, nurses, physiotherapists, pharmacists, and occupational therapists), limited support services (such as laboratory services or diagnostic imaging), and decreased availability of community services (including home care and social support services).[17] For example, in 1 large US heart failure registry, patients discharged on a weekend received less complete discharge instructions than those discharged on weekdays.[11] Given that early follow‐up postdischarge is associated with better outcomes,[18, 19] future studies should also explore whether patterns of patient follow‐up differ after weekend versus weekday discharges.

Although we were able to capture all interactions with the healthcare system in a single payer system with universal access, there are some limitations to our study. First, we used administrative data, which preclude fully adjusting for severity of diagnoses or functional status, although we used proxies such as admission from/discharge to a long‐term care facility.[20, 21] Second, we did not have access to process of care measures such as diagnostic testing or prescribing data, and thus cannot determine whether quality of care or patient adherence differed by the day of the week they were discharged on, although this seems unlikely. Third, although postdischarge follow‐up may be associated with better outcomes,[18, 19] we were unable to adjust for patterns of outpatient follow‐up in this study. Fourth, we acknowledge that death or readmission soon after discharge does not necessarily mean that the quality of care during the preceding hospitalization was suboptimal or that these deaths or readmissions were even potentially preventable. Many factors influence postdischarge mortality and/or readmission, and quality of inpatient care is only one.[22, 23, 24, 25] Fifth, although some may express concern that LOS may be a mediator in the causal pathway between discharge decision and postdischarge events, and that adjusting for LOS in analyses could thus spuriously obscure a true association, it is worth pointing out that our 2 sensitivity analyses to explore this (the 1 in which we excluded LOS from the analyses and the 1 in which we included expected LOS rather than the actual LOS) revealed nearly identical point estimates and 95% CI as our main analysis. Finally, as our study is observational, we cannot definitively conclude causality, nor can we exclude an 18% excess risk for patients discharged on weekends (or a 22% lower risk either), given our 95% CI for postdischarge adverse outcomes.

CONCLUSION

We found that the proportion of patients discharged on weekends is lower than the proportion admitted on weekends. We also found that lower risk/less severely ill patients appear to be preferentially discharged on weekends, and as a result, postdischarge outcomes are similar between weekend and weekday discharges despite shorter LOS and less availability of outpatient resources for patients discharged on a weekend. The reasons why more complicated patients are not discharged on weekends deserves further study, as safely increasing weekend discharge rates would improve efficiency and safety (by reducing unnecessary exposure to in‐hospital adverse events such as falls, unnecessary urinary catheterizations, and healthcare‐acquired infections). Although hospital admission has become a 24/7 business, we believe that hospital discharge processes should strive for the same level of efficiency.

Hospitals typically reduce staffing levels and the availability of diagnostic, laboratory, and treatment services on weekends, and patients admitted on weekends exhibit poorer in‐hospital outcomes for several medical conditions.[1, 2, 3, 4, 5, 6, 7, 8, 9] Whether or not patients discharged on weekends have worse clinical outcomes has been less well studied.[10, 11, 12] Discharge rates on Saturday and Sunday are lower than for the other 5 days of the week,[12] but bed shortages and hospital overcrowding have increased the demand for maximizing 24/7 week‐round discharge efficiency. Given that the number of patients discharged on weekends is likely to continue to increase, it is important to assess the risk of weekend discharge on outcomes monitored as performance indicators by organizations such as the Centers for Medicare and Medicaid Services, the American Medical Association Physicians Consortium for Performance Improvement, the National Quality Forum, and the Joint Commission.

Thus, we designed this study to evaluate baseline characteristics, length of stay (LOS), and postdischarge outcomes for general internal medicine (GIM) patients in teaching hospitals discharged on weekends compared to weekdays. Our objective was to determine whether postdischarge outcomes differed for patients discharged on weekends versus weekdays.

METHODS

RESULTS

Weekday Versus Weekend Discharge

Although patients admitted on weekdays and weekends were very similar (data available upon request), patients discharged on weekends (compared to those discharged on weekdays) were younger, more likely to be discharged home without additional support, and had fewer comorbidities (Table 1, Figure 1). Patients discharged on weekends had shorter lengths of stay than those discharged on weekdays (5.6 days vs 7.9 days, P<0.0001). In adjusted linear regression analyses, this 2.3‐day difference remained statistically significant (adjusted P value <0.0001).

Figure 1

Patients discharged on a weekend exhibited lower unadjusted 30‐day rates of death or readmission than those discharged on a weekday (10.6% vs 13.2%), but these differences disappeared after multivariable adjustment that accounted for differences in risk profile (aOR: 0.94, 95% CI: 0.771.16 (Table 2). Results were similar in sensitivity analyses adjusting for LACE scores without LOS included (aOR: 0.88, 95% CI: 0.711.08) or adjusting for LACE scores using expected LOS rather than actual LOS (aOR: 0.90, 95% CI: 0.731.10). Restricting the analysis to only those patients deemed to be at high risk for events due to LACE scores of 10 or greater confirmed that weekend and weekday discharges had similar outcomes in the first 30 days after discharge (aOR: 1.09, 95% CI: 0.851.41, Table 2). Similar patterns were seen when we included ED visits as part of the composite endpoint (ie, death, unplanned readmission, or unplanned ED visit within 30 days of discharge) (Table 2).

Table 2.Postdischarge Outcomes After a General Internal Medicine Hospitalization in a Teaching Hospital

	Weekend Discharge, n/N (%)	Weekday Discharge, n/N (%)	Unadjusted P Value	aOR* (95% CI)	Adjusted P Value
NOTE: Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; ED, emergency department; LACE, length of hospital stay, acuity of admission, comorbidity burden quantified using the Charlson Comorbidity Index, and emergency department visits in the 6 months prior to admission. *Multivariable models adjust for age, sex, hospital, and LACE score at time of discharge from index hospitalization. Weekday discharge is reference group for odds ratios.
Death/readmission within 30 days
All 7 teaching hospitals, all patients	121/1146 (10.6)	901/6845 (13.2)	0.01	0.94 (0.77‐1.16)	0.58
All 7 teaching hospitals, but only patients with LACE <10	37/647 (5.7)	225/2753 (8.2)	0.04	0.72 (0.50, 1.03)	0.07
All 7 teaching hospitals, but only patients with LACE 10	84/499 (16.8)	676/4092 (16.5)	0.86	1.09 (0.85‐1.41)	0.49
Death/readmission/ED visit within 30 days
All 7 teaching hospitals, all patients	218/1146 (19.0)	1445/6845 (21.1)	0.11	0.98 (0.83‐1.15)	0.79
All 7 teaching hospitals, but only patients with LACE <10	90/647 (13.9)	460/2753 (16.7)	0.08	0.83 (0.64‐1.06)	0.13
All 7 teaching hospitals, but only patients with LACE 10	128/499 (25.7)	985/4092 (24.1)	0.44	1.12 (0.90‐1.39)	0.31
Death within 30 days
All 7 teaching hospitals, all patients	24/1146 (2.1)	215/6845 (3.1)	0.05	0.97 (0.63‐1.51)	0.89
All 7 teaching hospitals, but only patients with LACE <10	4/647 (0.6)	23/2753 (0.8)	0.58	0.89 (0.30, 2.62)	0.83
All 7 teaching hospitals, but only patients with LACE 10	20/499 (4.0)	192/4092 (4.7)	0.49	0.99 (0.61‐1.61)	0.98
Readmission within 30 days
All 7 teaching hospitals, all patients	105/1146 (9.2)	751/6845 (11.0)	0.07	0.94 (0.76‐1.17)	0.59
All 7 teaching hospitals, but only patients with LACE <10	33/647 (5.1)	211/2753 (7.7)	0.02	0.68 (0.46‐0.99)	0.04
All 7 teaching hospitals, but only patients with LACE 10	72/499 (14.4)	540/4092 (13.2)	0.44	1.14 (0.87‐1.49)	0.34
ED visit within 30 days
All 7 teaching hospitals, all patients	182/1146 (15.9)	1118/6845 (16.3)	0.70	1.00 (0.84‐1.19)	0.99
All 7 teaching hospitals, but only patients with LACE <10	83/647 (12.8)	412/2753 (15.0)	0.17	0.84 (0.65, 1.09)	0.20
All 7 teaching hospitals, but only patients with LACE 10	99/499 (19.8)	706/4092 (17.3)	0.15	1.17 (0.92‐1.48)	0.20

DISCUSSION

Our data suggest that patients discharged from the GIM teaching wards we studied on weekends were appropriately triaged, as they did not exhibit a higher risk of adverse events postdischarge. Although patients discharged on weekends tended to be younger and had less comorbidities than those discharged during the week, we adjusted for baseline covariates in analyses, and we did not find an association between weekend discharge and increased postdischarge events even among the subset of patients deemed to be at high risk for postdischarge adverse events (based on high LACE scores). To our knowledge, although we previously examined this issue in patients with a most‐responsible diagnosis of heart failure,[10] examining weekend versus weekday discharges in the full gamut of general medical patients admitted to teaching hospitals has not previously been examined.

In our previous study[10] of over 24,000 heart failure patients discharged over 10 years (up to June 2009, therefore no overlap with any patients in this study), we also found that patients discharged on the weekends were younger, had fewer comorbidities, and shorter lengths of stay. Although postdischarge death/readmission rates were higher for weekend discharged patients in our earlier study (21.1% vs 19.5%, adjusted hazard ratio: 1.15, 95% CI: 1.061.25), it is worth noting that this was almost entirely driven by data from nonteaching hospitals and cardiology wards. Thus, it is important to reiterate that the findings in our current study are for GIM wards in teaching hospitals and may not be generalizable to less‐structured nonteaching settings.

Although we did not study physician decision making, our results suggest that physicians are incorporating discharge day into their discharge decision making. They may be selecting younger patients with less comorbidities for weekend discharges, or they may be delaying the discharges of older patients with more comorbidities for weekday discharges. Either is not surprising given the realities of weekend inpatient care: reduced staffing and frequent cross‐coverage (of physicians, nurses, physiotherapists, pharmacists, and occupational therapists), limited support services (such as laboratory services or diagnostic imaging), and decreased availability of community services (including home care and social support services).[17] For example, in 1 large US heart failure registry, patients discharged on a weekend received less complete discharge instructions than those discharged on weekdays.[11] Given that early follow‐up postdischarge is associated with better outcomes,[18, 19] future studies should also explore whether patterns of patient follow‐up differ after weekend versus weekday discharges.

Although we were able to capture all interactions with the healthcare system in a single payer system with universal access, there are some limitations to our study. First, we used administrative data, which preclude fully adjusting for severity of diagnoses or functional status, although we used proxies such as admission from/discharge to a long‐term care facility.[20, 21] Second, we did not have access to process of care measures such as diagnostic testing or prescribing data, and thus cannot determine whether quality of care or patient adherence differed by the day of the week they were discharged on, although this seems unlikely. Third, although postdischarge follow‐up may be associated with better outcomes,[18, 19] we were unable to adjust for patterns of outpatient follow‐up in this study. Fourth, we acknowledge that death or readmission soon after discharge does not necessarily mean that the quality of care during the preceding hospitalization was suboptimal or that these deaths or readmissions were even potentially preventable. Many factors influence postdischarge mortality and/or readmission, and quality of inpatient care is only one.[22, 23, 24, 25] Fifth, although some may express concern that LOS may be a mediator in the causal pathway between discharge decision and postdischarge events, and that adjusting for LOS in analyses could thus spuriously obscure a true association, it is worth pointing out that our 2 sensitivity analyses to explore this (the 1 in which we excluded LOS from the analyses and the 1 in which we included expected LOS rather than the actual LOS) revealed nearly identical point estimates and 95% CI as our main analysis. Finally, as our study is observational, we cannot definitively conclude causality, nor can we exclude an 18% excess risk for patients discharged on weekends (or a 22% lower risk either), given our 95% CI for postdischarge adverse outcomes.

CONCLUSION

We found that the proportion of patients discharged on weekends is lower than the proportion admitted on weekends. We also found that lower risk/less severely ill patients appear to be preferentially discharged on weekends, and as a result, postdischarge outcomes are similar between weekend and weekday discharges despite shorter LOS and less availability of outpatient resources for patients discharged on a weekend. The reasons why more complicated patients are not discharged on weekends deserves further study, as safely increasing weekend discharge rates would improve efficiency and safety (by reducing unnecessary exposure to in‐hospital adverse events such as falls, unnecessary urinary catheterizations, and healthcare‐acquired infections). Although hospital admission has become a 24/7 business, we believe that hospital discharge processes should strive for the same level of efficiency.

Hospitals typically reduce staffing levels and the availability of diagnostic, laboratory, and treatment services on weekends, and patients admitted on weekends exhibit poorer in‐hospital outcomes for several medical conditions.[1, 2, 3, 4, 5, 6, 7, 8, 9] Whether or not patients discharged on weekends have worse clinical outcomes has been less well studied.[10, 11, 12] Discharge rates on Saturday and Sunday are lower than for the other 5 days of the week,[12] but bed shortages and hospital overcrowding have increased the demand for maximizing 24/7 week‐round discharge efficiency. Given that the number of patients discharged on weekends is likely to continue to increase, it is important to assess the risk of weekend discharge on outcomes monitored as performance indicators by organizations such as the Centers for Medicare and Medicaid Services, the American Medical Association Physicians Consortium for Performance Improvement, the National Quality Forum, and the Joint Commission.

Thus, we designed this study to evaluate baseline characteristics, length of stay (LOS), and postdischarge outcomes for general internal medicine (GIM) patients in teaching hospitals discharged on weekends compared to weekdays. Our objective was to determine whether postdischarge outcomes differed for patients discharged on weekends versus weekdays.

METHODS

RESULTS

Weekday Versus Weekend Discharge

Although patients admitted on weekdays and weekends were very similar (data available upon request), patients discharged on weekends (compared to those discharged on weekdays) were younger, more likely to be discharged home without additional support, and had fewer comorbidities (Table 1, Figure 1). Patients discharged on weekends had shorter lengths of stay than those discharged on weekdays (5.6 days vs 7.9 days, P<0.0001). In adjusted linear regression analyses, this 2.3‐day difference remained statistically significant (adjusted P value <0.0001).

Figure 1

Patients discharged on a weekend exhibited lower unadjusted 30‐day rates of death or readmission than those discharged on a weekday (10.6% vs 13.2%), but these differences disappeared after multivariable adjustment that accounted for differences in risk profile (aOR: 0.94, 95% CI: 0.771.16 (Table 2). Results were similar in sensitivity analyses adjusting for LACE scores without LOS included (aOR: 0.88, 95% CI: 0.711.08) or adjusting for LACE scores using expected LOS rather than actual LOS (aOR: 0.90, 95% CI: 0.731.10). Restricting the analysis to only those patients deemed to be at high risk for events due to LACE scores of 10 or greater confirmed that weekend and weekday discharges had similar outcomes in the first 30 days after discharge (aOR: 1.09, 95% CI: 0.851.41, Table 2). Similar patterns were seen when we included ED visits as part of the composite endpoint (ie, death, unplanned readmission, or unplanned ED visit within 30 days of discharge) (Table 2).

Table 2.Postdischarge Outcomes After a General Internal Medicine Hospitalization in a Teaching Hospital

	Weekend Discharge, n/N (%)	Weekday Discharge, n/N (%)	Unadjusted P Value	aOR* (95% CI)	Adjusted P Value
NOTE: Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; ED, emergency department; LACE, length of hospital stay, acuity of admission, comorbidity burden quantified using the Charlson Comorbidity Index, and emergency department visits in the 6 months prior to admission. *Multivariable models adjust for age, sex, hospital, and LACE score at time of discharge from index hospitalization. Weekday discharge is reference group for odds ratios.
Death/readmission within 30 days
All 7 teaching hospitals, all patients	121/1146 (10.6)	901/6845 (13.2)	0.01	0.94 (0.77‐1.16)	0.58
All 7 teaching hospitals, but only patients with LACE <10	37/647 (5.7)	225/2753 (8.2)	0.04	0.72 (0.50, 1.03)	0.07
All 7 teaching hospitals, but only patients with LACE 10	84/499 (16.8)	676/4092 (16.5)	0.86	1.09 (0.85‐1.41)	0.49
Death/readmission/ED visit within 30 days
All 7 teaching hospitals, all patients	218/1146 (19.0)	1445/6845 (21.1)	0.11	0.98 (0.83‐1.15)	0.79
All 7 teaching hospitals, but only patients with LACE <10	90/647 (13.9)	460/2753 (16.7)	0.08	0.83 (0.64‐1.06)	0.13
All 7 teaching hospitals, but only patients with LACE 10	128/499 (25.7)	985/4092 (24.1)	0.44	1.12 (0.90‐1.39)	0.31
Death within 30 days
All 7 teaching hospitals, all patients	24/1146 (2.1)	215/6845 (3.1)	0.05	0.97 (0.63‐1.51)	0.89
All 7 teaching hospitals, but only patients with LACE <10	4/647 (0.6)	23/2753 (0.8)	0.58	0.89 (0.30, 2.62)	0.83
All 7 teaching hospitals, but only patients with LACE 10	20/499 (4.0)	192/4092 (4.7)	0.49	0.99 (0.61‐1.61)	0.98
Readmission within 30 days
All 7 teaching hospitals, all patients	105/1146 (9.2)	751/6845 (11.0)	0.07	0.94 (0.76‐1.17)	0.59
All 7 teaching hospitals, but only patients with LACE <10	33/647 (5.1)	211/2753 (7.7)	0.02	0.68 (0.46‐0.99)	0.04
All 7 teaching hospitals, but only patients with LACE 10	72/499 (14.4)	540/4092 (13.2)	0.44	1.14 (0.87‐1.49)	0.34
ED visit within 30 days
All 7 teaching hospitals, all patients	182/1146 (15.9)	1118/6845 (16.3)	0.70	1.00 (0.84‐1.19)	0.99
All 7 teaching hospitals, but only patients with LACE <10	83/647 (12.8)	412/2753 (15.0)	0.17	0.84 (0.65, 1.09)	0.20
All 7 teaching hospitals, but only patients with LACE 10	99/499 (19.8)	706/4092 (17.3)	0.15	1.17 (0.92‐1.48)	0.20

DISCUSSION

Our data suggest that patients discharged from the GIM teaching wards we studied on weekends were appropriately triaged, as they did not exhibit a higher risk of adverse events postdischarge. Although patients discharged on weekends tended to be younger and had less comorbidities than those discharged during the week, we adjusted for baseline covariates in analyses, and we did not find an association between weekend discharge and increased postdischarge events even among the subset of patients deemed to be at high risk for postdischarge adverse events (based on high LACE scores). To our knowledge, although we previously examined this issue in patients with a most‐responsible diagnosis of heart failure,[10] examining weekend versus weekday discharges in the full gamut of general medical patients admitted to teaching hospitals has not previously been examined.

In our previous study[10] of over 24,000 heart failure patients discharged over 10 years (up to June 2009, therefore no overlap with any patients in this study), we also found that patients discharged on the weekends were younger, had fewer comorbidities, and shorter lengths of stay. Although postdischarge death/readmission rates were higher for weekend discharged patients in our earlier study (21.1% vs 19.5%, adjusted hazard ratio: 1.15, 95% CI: 1.061.25), it is worth noting that this was almost entirely driven by data from nonteaching hospitals and cardiology wards. Thus, it is important to reiterate that the findings in our current study are for GIM wards in teaching hospitals and may not be generalizable to less‐structured nonteaching settings.

Although we did not study physician decision making, our results suggest that physicians are incorporating discharge day into their discharge decision making. They may be selecting younger patients with less comorbidities for weekend discharges, or they may be delaying the discharges of older patients with more comorbidities for weekday discharges. Either is not surprising given the realities of weekend inpatient care: reduced staffing and frequent cross‐coverage (of physicians, nurses, physiotherapists, pharmacists, and occupational therapists), limited support services (such as laboratory services or diagnostic imaging), and decreased availability of community services (including home care and social support services).[17] For example, in 1 large US heart failure registry, patients discharged on a weekend received less complete discharge instructions than those discharged on weekdays.[11] Given that early follow‐up postdischarge is associated with better outcomes,[18, 19] future studies should also explore whether patterns of patient follow‐up differ after weekend versus weekday discharges.

Although we were able to capture all interactions with the healthcare system in a single payer system with universal access, there are some limitations to our study. First, we used administrative data, which preclude fully adjusting for severity of diagnoses or functional status, although we used proxies such as admission from/discharge to a long‐term care facility.[20, 21] Second, we did not have access to process of care measures such as diagnostic testing or prescribing data, and thus cannot determine whether quality of care or patient adherence differed by the day of the week they were discharged on, although this seems unlikely. Third, although postdischarge follow‐up may be associated with better outcomes,[18, 19] we were unable to adjust for patterns of outpatient follow‐up in this study. Fourth, we acknowledge that death or readmission soon after discharge does not necessarily mean that the quality of care during the preceding hospitalization was suboptimal or that these deaths or readmissions were even potentially preventable. Many factors influence postdischarge mortality and/or readmission, and quality of inpatient care is only one.[22, 23, 24, 25] Fifth, although some may express concern that LOS may be a mediator in the causal pathway between discharge decision and postdischarge events, and that adjusting for LOS in analyses could thus spuriously obscure a true association, it is worth pointing out that our 2 sensitivity analyses to explore this (the 1 in which we excluded LOS from the analyses and the 1 in which we included expected LOS rather than the actual LOS) revealed nearly identical point estimates and 95% CI as our main analysis. Finally, as our study is observational, we cannot definitively conclude causality, nor can we exclude an 18% excess risk for patients discharged on weekends (or a 22% lower risk either), given our 95% CI for postdischarge adverse outcomes.

CONCLUSION

We found that the proportion of patients discharged on weekends is lower than the proportion admitted on weekends. We also found that lower risk/less severely ill patients appear to be preferentially discharged on weekends, and as a result, postdischarge outcomes are similar between weekend and weekday discharges despite shorter LOS and less availability of outpatient resources for patients discharged on a weekend. The reasons why more complicated patients are not discharged on weekends deserves further study, as safely increasing weekend discharge rates would improve efficiency and safety (by reducing unnecessary exposure to in‐hospital adverse events such as falls, unnecessary urinary catheterizations, and healthcare‐acquired infections). Although hospital admission has become a 24/7 business, we believe that hospital discharge processes should strive for the same level of efficiency.

A 40‐year‐old Sudanese man was admitted due to worsening abdominal pain with recurrent ascites. He had a history of hepatitis B (HBV) infection and diabetes. He previously drank 3 beers per day on the weekends, but he had not consumed alcohol in over a year. He was born in Sudan but lived in Egypt most of his adult life; he immigrated to the United States 6 years previously. He was hospitalized out of state 9 months ago for a swollen abdomen and underwent an exploratory laparotomy that reportedly was unremarkable except for ascites.

Portal hypertension due to liver disease is the most common cause of ascites. This patient has a known risk factor for liver disease (history of HBV infection). Although his reported alcohol consumption is low, there is a synergistic effect on liver injury in the setting of chronic hepatitis. Abdominal pain in the setting of ascites needs to be urgently evaluated to exclude spontaneous bacterial peritonitis (SBP). Also, because chronic HBV infection is the major risk factor for hepatocellular carcinoma in the world, malignant ascites is in the differential. Hepatic vascular thrombosis and tuberculous peritonitis (given the patient's country of origin and travel history) also should be considered. The most appropriate initial test would be a diagnostic paracentesis to support or exclude the presence of SBP and direct the evaluation toward liver disease or other less‐common causes of ascites.

The patient was seen as an outpatient 5 months prior to admission with transient fever and joint pains. Laboratory studies at that visit were notable for a serum albumin of 3.2 g/dL (normal 3.55), 2.4 g of predicted 24‐hour protein on urinalysis (normal <30 mg per 24 hours), creatinine of 0.5 mg/dL (normal 0.81.3), and a positive hepatitis B surface antibody. The working diagnosis was a nonspecific viral syndrome and his symptoms resolved without treatment. One month later, he developed ascites and mild lower extremity edema. Additional laboratory studies at that time showed a normocytic anemia with hemoglobin 11.7 g/dL (normal 13.517.5) and leukopenia with white blood cell count of 2.4 10⁹/L (normal 3.510.5), neutrophil count of 1.45 10⁹/L (normal 1.77.0), and lymphocyte count of 0.58 10⁹/L (normal 0.902.90). Transaminases, serum bilirubin, prothrombin time, alpha fetoprotein, and peripheral blood smear were normal. Human immunodeficiency virus antibody screen and QuantiFERON‐TB assay were negative. Hemoglobin A1c was 6.2% (normal 4.06.0). Repeat urinalysis demonstrated 883 mg of predicted 24‐hour protein. Computed tomography (CT) of the abdomen showed a large amount of intra‐abdominal ascites; the liver and spleen were normal, and there were no varices or other evidence of portal hypertension. Echocardiogram was normal except for a small inferior vena cava (IVC) and a mildly increased right ventricular systolic pressure of 32 mm Hg (systolic blood pressure 98 mm Hg). Due to the indeterminate cause for the patient's ascites, referral was made for gastroenterology evaluation with consideration for a paracentesis.

Cirrhotic ascites seems less likely. Postsinusoidal causes of portal hypertension (eg, cardiomyopathy) are also less likely given the absence of suggestive findings on echocardiography. Malignant ascites also appears less probable in the absence of suggestive findings such as mass lesions, lymphadenopathy, or peritoneal carcinomatosis on CT imaging. The suspicion for tuberculous peritonitis is lower with the negative QuantiFERON‐TB test. Hypoalbuminemia, normocytic anemia, leukopenia, and proteinuria all suggest a systemic inflammatory condition (eg, systemic lupus erythematosus [SLE]) with inflammatory serositis causing ascites). Nephrotic syndrome can cause hypoalbuminemia, edema, and ascites, but his total urine protein losses of <3.5 grams per 24 hours are not in keeping with this diagnosis. Other uncommon causes of ascites such as chylous ascites have not yet been excluded. The most appropriate next step remains ascitic fluid analysis.

A paracentesis yielded 7.8 L of clear‐yellow fluid and improvement in his abdominal discomfort. Analysis showed 224 total nucleated cells/L with 2% neutrophils, 57% lymphocytes, and 37% monocytes. Ascites total protein was 3.8 g/dL and glucose was 55 mg/dL. Gram stain and culture were negative, and cytology was negative for malignancy but showed lymphocytes, plasma cells, monocytes, and reactive mesothelial cells interpreted as consistent with chronic inflammation. The serum‐ascites albumin gradient (SAAG) was not obtained.

With a low leukocyte count and a paucity of neutrophils, this is not SBP. The ascites fluid did not have a chylous appearance. The SAAG, which can distinguish between portal hypertensive and nonportal hypertensive causes for ascites using a cutoff of 1.1 g/dL, was not done. The total protein was high, arguing against cirrhosis. High protein ascites with a high SAAG would suggest a posthepatic source of portal hypertension (eg, Budd‐Chiari syndrome, constrictive pericarditis). High protein ascites with a low SAAG would suggest an inflammatory or malignant source of ascites. The relative lymphocytosis in the ascites fluid suggests an inflammatory process, but is a nonspecific finding. The negative cytology does not completely exclude a malignancy, but given the absence of findings on the CT, malignant ascites is less likely.

Three months before admission, the patient underwent a repeat large‐volume paracentesis and a liver biopsy. The biopsy showed ectopic portal vein branches consistent with hepatoportal sclerosis, but no actual sclerosis was identified. The pathologist concluded that the findings suggested noncirrhotic portal hypertension due to a vascular in‐flow abnormality. Abdominal ultrasound with Doppler was unremarkable other than slightly increased echogenicity of the liver. Magnetic resonance (MR) angiogram showed narrowing of the intra‐abdominal IVC at the level of the diaphragm. Because of concern that hepatic congestion from high pressures in the narrowed IVC was leading to poor vascular inflow as suggested by the biopsy findings, an inferior vena cavagram was performed. This study was normal, although no transhepatic pressure measurements were obtained. Three stool specimens and 2 urine specimens were negative for parasites. The patient required repeat large‐volume paracenteses monthly. SBP was again ruled out, but no other diagnostic labs were obtained. He had anorexia with poor oral intake each time his abdomen became distended.

The patient was started on furosemide 1 month prior to admission to the hospital but had only a slight improvement in the ascites. His other medications included insulin, tamsulosin, and hydrocodone‐acetaminophen. Five days prior to admission, he underwent a diagnostic laparoscopy, which showed only ascites and small adhesions to the anterior abdominal wall. There was no visual evidence of malignancy, and the surgeon commented that the liver was normal. No additional biopsies were obtained.

The liver biopsy findings could be seen in noncirrhotic portal hypertension, although this diagnosis would be unlikely without splenomegaly, varices, or other signs of portal hypertension. However, 2 possible etiologies for noncirrhotic portal hypertension in this patient would be hepatic congestion from the narrowed IVC (although the normal IVC study argues against this) and hepatic schistosomiasis. Schistosomiasis is an important cause of noncirrhotic portal hypertension in endemic areas like this patient's country of origin, but the negative stool and urine studies, combined with the lack of granulomas or fibrosis seen on biopsy, make this condition unlikely.

Systemic amyloidosis (primary or secondary) could also be a cause of ascites and could present with multiorgan involvement (diarrhea and nephrotic syndrome). Amyloid deposits would have probably been seen in the liver biopsy, if present, but may not have been apparent unless specific stains (Congo red) were performed.

Evaluation for systemic, inflammatory autoimmune processes is indicated. Serum autoantibodies (anti‐nuclear antibody [ANA] and extractable nuclear antigens), and a serum and 24‐hour urine protein electrophoresis would be appropriate diagnostic tests. Peritoneal biopsies would have been helpful to assess for serosal diseases.

The patient subsequently developed acute right‐sided abdominal pain requiring urgent evaluation and admission to the hospital. He was initially assessed by a general surgeon, who found no evidence of postoperative complications. His temperature was 36.7C, blood pressure 105/64, heart rate 82, respiratory rate 16, and oxygen saturation 97% on room air. He appeared chronically ill, but he was in no distress and he had a normal mental status. Cardiac exam was normal except for mild jugular venous distension. He had mild bibasilar lung crackles. His abdomen was distended with superficial abdominal tenderness and a fluid wave, but he had normal bowel sounds and no peritoneal signs. He had mild scrotal edema but no peripheral edema. Joint exam did not suggest synovitis and there were no rashes or oral ulcers. Lactate was 0.9 mmol/L (normal 0.62.3), albumin was 2.6 g/dL, and prealbumin was 9 mg/dL (normal 1938). Erythrocyte sedimentation rate and C‐reactive protein were 46 mm/hour (normal <22) and 33.1 mg/L (normal 8), respectively. He had a normocytic anemia and leukopenia. Liver tests and routine chemistries were normal. Serum protein electrophoresis indicated no monoclonal protein. Complete 24‐hour urine collection showed 1.2 g of protein (normal <102 mg). Paracentesis of 3.4 L demonstrated 227 total nucleated cells/L with 2% neutrophils. Following the fluid removal, he had improvement in his pain, which he felt was related to the ascites rather than the recent surgery. Ascites total protein was 3.9 g/dL and ascites albumin was 1.7 g/dL. Ascites culture was negative for infection. Serum Schistosoma immunoglobulin G (IgG) antibody was positive at 3.53 (normal <1.00).

Further history revealed prior episodes of polyarticular joint pain and swelling in his hands and knees 5 years before admission. At that time, he reported a diffuse, pruritic, papular body rash. In addition, he noticed that his fingertips and toes turned white with cold exposure.

Importantly, surgical and infectious complications have been excluded. High protein ascites with a low SAAG of 0.9 suggests an inflammatory source of ascites. The follow‐up clinical data (arthritis, normocytic anemia, leukopenia, rash, Raynaud's phenomenon) suggest a systemic inflammatory syndrome such as SLE, with accompanying serositis. Serologic testing for autoantibodies would be recommended. Peritoneal biopsies, if obtained, may have demonstrated chronic, inflammatory infiltrate (nonspecific) or leukocytoclastic vasculitis (strongly supportive).

ANA enzyme immunoassay was >12 U (normal 1.0 U). Extractable nuclear antigens revealed positive autoantibodies for anti‐SSA, anti‐SSB, and anti‐ribosomal P. Moreover, double‐stranded DNA IgG antibody was 120 IU/mL (normal <30 IU/mL) and C3, C4, and total complement levels were low.

The clinical data support a diagnosis of SLE with serositis. Treatment of the underlying connective tissue disease will typically result in resolution of the ascites; diuretic therapy is generally ineffective.

In consultation with rheumatology and gastroenterology specialists, the diagnosis of SLE was made based on criteria of serositis, persistent leukopenia, arthritis, renal disease (proteinuria), positive ANA, elevated ds‐DNA antibodies, and hypocomplementemia. MR imaging of the abdominal vasculature demonstrated no evidence of vasculitis. The patient was given intravenous methylprednisolone 1 g daily for 3 days followed by high‐dose oral corticosteroids with a gradual taper. He was also started on mycophenolate mofetil as a steroid‐sparing medication (which was later changed to leflunomide due to persistent leukopenia) and hydroxychloroquine. His isolated positive Schistosoma IgG antibody in the absence of other findings was consistent with past exposure or infection. The infectious disease specialist felt there was no evidence of active schistosomiasis, but recommended treatment with a single dose of praziquantel due to the potential benefit with low risk of side effects. The patient had ongoing improvement following dismissal. He had 1 additional paracentesis of 4.1 L, 10 days after his hospitalization, and his ascites and proteinuria resolved. At the 5‐year follow‐up visit, there had been no recurrence of abdominal ascites or abdominal pain. He remains on low‐dose prednisone at 5 mg daily, leflunomide, and hydroxychloroquine.

COMMENTARY

This patient had recurrent ascites with 29.6 L removed over the 4 months prior to admission and an additional 3.4 L during his hospitalization. His outpatient providers initially considered a portal hypertensive etiology of his ascites due to his history of HBV and prior alcohol use. They also appropriately investigated for a possible infectious process. They next directed their evaluation toward the liver biopsy findings, which raised concern for a vascular inflow abnormality. However, the evaluation could have been performed more rapidly and far more cost‐efficiently had a diagnostic paracentesis with calculation of the SAAG been performed early in the evaluation.

The SAAG, which was first described in 1983 by Par and colleagues, is a parameter reflecting the oncotic pressure gradient between the vascular bed and the interstitial splanchnic or ascitic fluid. [1] In the classic study by Runyon and colleagues, a SAAG difference of 1.1 g/dL correctly differentiated causes of ascites due to portal hypertension from those that were not due to portal hypertension 96.7% of the time. [2] Conditions such as nephrotic syndrome, peritoneal carcinomatosis, and serositis (lupus peritonitis) can cause ascites in patients without portal hypertension.

Serositis in the form of pleuritis and/or pericarditis is a common feature of SLE, and ascites has been described in 8% to 11% of SLE patients.[3] However, massive ascites due to lupus peritonitis as a presenting symptom is rare.[4] More common causes of ascites in the setting of SLE include nephrotic syndrome, heart failure, protein‐losing enteropathy, constrictive pericarditis, Budd‐Chiari syndrome, indolent infections such as tuberculosis, and chylous ascites.[5, 6, 7] Of note, lupus peritonitis may be chronic or acute. Chronic ascites develops insidiously with few manifestations of active lupus and may be painless, whereas ascites from acute lupus peritonitis typically develops rapidly and presents with acute abdominal pain and other signs of increased lupus activity.[3, 5, 6, 8, 9]

Ascites from lupus peritonitis may be due to marked serosal exudative accumulation with reduced absorptive capacity in the peritoneum.[3, 4, 10] Other possible causes include peritoneal inflammation from deposition of immune complexes or vasculitis of peritoneal vessels and visceral serous membranes.[4, 9, 11] Although subserosal and submucosal vasculitis have been found in acute ascites, chronic ascites may be related to scarring from vasculitis and serosal inflammation leading to poor venous and lymph drainage.[9] Ascitic fluid characteristics from lupus peritonitis include a SAAG <1.1, presence of white blood cells anywhere in a broad range from 10 to 1630/L, and a range of fluid protein from 3.4 to 4.7 mg/dL.[3] Although not tested in this patient, findings of low complement levels, positive ANA, and elevated anti‐DNA antibody in the ascitic fluid would be supportive of lupus peritonitis, but not specific.[5, 9, 12] Lupus erythematosus cells are occasionally found in the ascitic fluid, but do not rule out other causes of ascites.[9] On retrospective analysis, lupus erythematosus cells were not seen in this patient's pathology specimens.

Treatment of lupus peritonitis and ascites is with high‐dose glucocorticoid therapy, but many patients may need a second immunosuppressant, possibly because of impaired peritoneal circulation from chronic inflammation leading to decreased drug delivery.[13, 14] Chronic ascites may be recalcitrant to systemic glucocorticoids,[3] so a possible alternative therapy is intraperitoneal injection of triamcinolone, which successfully treated massive ascites in a patient who did not respond to oral glucocorticoid treatment.[13] Although ascites may be refractory in some patients, those with chronic lupus peritonitis can generally achieve remission, yet the overall prognosis depends on the presence and severity of multiorgan involvement from SLE. As with any SLE patient, there are also risks of infection from immunosuppression and increased cardiovascular risks.

This patient's evaluation and treatment could have been expedited if he had undergone a paracenteses with determination of the SAAG early in his workup. It is not known why the SAAG was not obtained despite multiple outpatient visits and paracenteses, his history of HBV, and prior alcohol use. This may have been simply an unfortunate oversight. Alternatively, it may have been that his outpatient providers focused on tantalizing clues such as his country of origin, which led to concern for schistosomiasis, and the biopsy findings suggestive of a vascular inflow abnormality that led to further extensive testing. In so doing, the clinicians committed several diagnostic errors, including multiple alternatives bias, anchoring, and confirmation bias.[15] As a result, the patient accrued excess charges of $64,000 from multiple tests, laparoscopic surgery, and 2 hospitalizations. This case highlights how cognitive errors introduce costly variability into patient care, especially when a simple and accurate test is at the beginning of the decision tree.

CLINICAL TEACHING POINTS

1. Diagnostic paracentesis, with calculation of the serum‐ascites albumin gradient, should be the first test in the workup for ascites and can distinguish portal hypertensive causes from nonportal hypertensive causes.
2. Ascites related to SLE can be acute or chronic and caused by bowel infarction, perforation, pancreatitis, mesenteric vasculitis, nephrotic syndrome, heart failure, protein‐losing enteropathy, constrictive pericarditis, lupus peritonitis, Budd‐Chiari syndrome, or serositis (lupus peritonitis).
3. Ascites caused by lupus peritonitis is rare. Once treated, management should be directed toward keeping the SLE in remission.

A 40‐year‐old Sudanese man was admitted due to worsening abdominal pain with recurrent ascites. He had a history of hepatitis B (HBV) infection and diabetes. He previously drank 3 beers per day on the weekends, but he had not consumed alcohol in over a year. He was born in Sudan but lived in Egypt most of his adult life; he immigrated to the United States 6 years previously. He was hospitalized out of state 9 months ago for a swollen abdomen and underwent an exploratory laparotomy that reportedly was unremarkable except for ascites.

Portal hypertension due to liver disease is the most common cause of ascites. This patient has a known risk factor for liver disease (history of HBV infection). Although his reported alcohol consumption is low, there is a synergistic effect on liver injury in the setting of chronic hepatitis. Abdominal pain in the setting of ascites needs to be urgently evaluated to exclude spontaneous bacterial peritonitis (SBP). Also, because chronic HBV infection is the major risk factor for hepatocellular carcinoma in the world, malignant ascites is in the differential. Hepatic vascular thrombosis and tuberculous peritonitis (given the patient's country of origin and travel history) also should be considered. The most appropriate initial test would be a diagnostic paracentesis to support or exclude the presence of SBP and direct the evaluation toward liver disease or other less‐common causes of ascites.

The patient was seen as an outpatient 5 months prior to admission with transient fever and joint pains. Laboratory studies at that visit were notable for a serum albumin of 3.2 g/dL (normal 3.55), 2.4 g of predicted 24‐hour protein on urinalysis (normal <30 mg per 24 hours), creatinine of 0.5 mg/dL (normal 0.81.3), and a positive hepatitis B surface antibody. The working diagnosis was a nonspecific viral syndrome and his symptoms resolved without treatment. One month later, he developed ascites and mild lower extremity edema. Additional laboratory studies at that time showed a normocytic anemia with hemoglobin 11.7 g/dL (normal 13.517.5) and leukopenia with white blood cell count of 2.4 10⁹/L (normal 3.510.5), neutrophil count of 1.45 10⁹/L (normal 1.77.0), and lymphocyte count of 0.58 10⁹/L (normal 0.902.90). Transaminases, serum bilirubin, prothrombin time, alpha fetoprotein, and peripheral blood smear were normal. Human immunodeficiency virus antibody screen and QuantiFERON‐TB assay were negative. Hemoglobin A1c was 6.2% (normal 4.06.0). Repeat urinalysis demonstrated 883 mg of predicted 24‐hour protein. Computed tomography (CT) of the abdomen showed a large amount of intra‐abdominal ascites; the liver and spleen were normal, and there were no varices or other evidence of portal hypertension. Echocardiogram was normal except for a small inferior vena cava (IVC) and a mildly increased right ventricular systolic pressure of 32 mm Hg (systolic blood pressure 98 mm Hg). Due to the indeterminate cause for the patient's ascites, referral was made for gastroenterology evaluation with consideration for a paracentesis.

Cirrhotic ascites seems less likely. Postsinusoidal causes of portal hypertension (eg, cardiomyopathy) are also less likely given the absence of suggestive findings on echocardiography. Malignant ascites also appears less probable in the absence of suggestive findings such as mass lesions, lymphadenopathy, or peritoneal carcinomatosis on CT imaging. The suspicion for tuberculous peritonitis is lower with the negative QuantiFERON‐TB test. Hypoalbuminemia, normocytic anemia, leukopenia, and proteinuria all suggest a systemic inflammatory condition (eg, systemic lupus erythematosus [SLE]) with inflammatory serositis causing ascites). Nephrotic syndrome can cause hypoalbuminemia, edema, and ascites, but his total urine protein losses of <3.5 grams per 24 hours are not in keeping with this diagnosis. Other uncommon causes of ascites such as chylous ascites have not yet been excluded. The most appropriate next step remains ascitic fluid analysis.

A paracentesis yielded 7.8 L of clear‐yellow fluid and improvement in his abdominal discomfort. Analysis showed 224 total nucleated cells/L with 2% neutrophils, 57% lymphocytes, and 37% monocytes. Ascites total protein was 3.8 g/dL and glucose was 55 mg/dL. Gram stain and culture were negative, and cytology was negative for malignancy but showed lymphocytes, plasma cells, monocytes, and reactive mesothelial cells interpreted as consistent with chronic inflammation. The serum‐ascites albumin gradient (SAAG) was not obtained.

With a low leukocyte count and a paucity of neutrophils, this is not SBP. The ascites fluid did not have a chylous appearance. The SAAG, which can distinguish between portal hypertensive and nonportal hypertensive causes for ascites using a cutoff of 1.1 g/dL, was not done. The total protein was high, arguing against cirrhosis. High protein ascites with a high SAAG would suggest a posthepatic source of portal hypertension (eg, Budd‐Chiari syndrome, constrictive pericarditis). High protein ascites with a low SAAG would suggest an inflammatory or malignant source of ascites. The relative lymphocytosis in the ascites fluid suggests an inflammatory process, but is a nonspecific finding. The negative cytology does not completely exclude a malignancy, but given the absence of findings on the CT, malignant ascites is less likely.

Three months before admission, the patient underwent a repeat large‐volume paracentesis and a liver biopsy. The biopsy showed ectopic portal vein branches consistent with hepatoportal sclerosis, but no actual sclerosis was identified. The pathologist concluded that the findings suggested noncirrhotic portal hypertension due to a vascular in‐flow abnormality. Abdominal ultrasound with Doppler was unremarkable other than slightly increased echogenicity of the liver. Magnetic resonance (MR) angiogram showed narrowing of the intra‐abdominal IVC at the level of the diaphragm. Because of concern that hepatic congestion from high pressures in the narrowed IVC was leading to poor vascular inflow as suggested by the biopsy findings, an inferior vena cavagram was performed. This study was normal, although no transhepatic pressure measurements were obtained. Three stool specimens and 2 urine specimens were negative for parasites. The patient required repeat large‐volume paracenteses monthly. SBP was again ruled out, but no other diagnostic labs were obtained. He had anorexia with poor oral intake each time his abdomen became distended.

The patient was started on furosemide 1 month prior to admission to the hospital but had only a slight improvement in the ascites. His other medications included insulin, tamsulosin, and hydrocodone‐acetaminophen. Five days prior to admission, he underwent a diagnostic laparoscopy, which showed only ascites and small adhesions to the anterior abdominal wall. There was no visual evidence of malignancy, and the surgeon commented that the liver was normal. No additional biopsies were obtained.

The liver biopsy findings could be seen in noncirrhotic portal hypertension, although this diagnosis would be unlikely without splenomegaly, varices, or other signs of portal hypertension. However, 2 possible etiologies for noncirrhotic portal hypertension in this patient would be hepatic congestion from the narrowed IVC (although the normal IVC study argues against this) and hepatic schistosomiasis. Schistosomiasis is an important cause of noncirrhotic portal hypertension in endemic areas like this patient's country of origin, but the negative stool and urine studies, combined with the lack of granulomas or fibrosis seen on biopsy, make this condition unlikely.

Systemic amyloidosis (primary or secondary) could also be a cause of ascites and could present with multiorgan involvement (diarrhea and nephrotic syndrome). Amyloid deposits would have probably been seen in the liver biopsy, if present, but may not have been apparent unless specific stains (Congo red) were performed.

Evaluation for systemic, inflammatory autoimmune processes is indicated. Serum autoantibodies (anti‐nuclear antibody [ANA] and extractable nuclear antigens), and a serum and 24‐hour urine protein electrophoresis would be appropriate diagnostic tests. Peritoneal biopsies would have been helpful to assess for serosal diseases.

The patient subsequently developed acute right‐sided abdominal pain requiring urgent evaluation and admission to the hospital. He was initially assessed by a general surgeon, who found no evidence of postoperative complications. His temperature was 36.7C, blood pressure 105/64, heart rate 82, respiratory rate 16, and oxygen saturation 97% on room air. He appeared chronically ill, but he was in no distress and he had a normal mental status. Cardiac exam was normal except for mild jugular venous distension. He had mild bibasilar lung crackles. His abdomen was distended with superficial abdominal tenderness and a fluid wave, but he had normal bowel sounds and no peritoneal signs. He had mild scrotal edema but no peripheral edema. Joint exam did not suggest synovitis and there were no rashes or oral ulcers. Lactate was 0.9 mmol/L (normal 0.62.3), albumin was 2.6 g/dL, and prealbumin was 9 mg/dL (normal 1938). Erythrocyte sedimentation rate and C‐reactive protein were 46 mm/hour (normal <22) and 33.1 mg/L (normal 8), respectively. He had a normocytic anemia and leukopenia. Liver tests and routine chemistries were normal. Serum protein electrophoresis indicated no monoclonal protein. Complete 24‐hour urine collection showed 1.2 g of protein (normal <102 mg). Paracentesis of 3.4 L demonstrated 227 total nucleated cells/L with 2% neutrophils. Following the fluid removal, he had improvement in his pain, which he felt was related to the ascites rather than the recent surgery. Ascites total protein was 3.9 g/dL and ascites albumin was 1.7 g/dL. Ascites culture was negative for infection. Serum Schistosoma immunoglobulin G (IgG) antibody was positive at 3.53 (normal <1.00).

Further history revealed prior episodes of polyarticular joint pain and swelling in his hands and knees 5 years before admission. At that time, he reported a diffuse, pruritic, papular body rash. In addition, he noticed that his fingertips and toes turned white with cold exposure.

Importantly, surgical and infectious complications have been excluded. High protein ascites with a low SAAG of 0.9 suggests an inflammatory source of ascites. The follow‐up clinical data (arthritis, normocytic anemia, leukopenia, rash, Raynaud's phenomenon) suggest a systemic inflammatory syndrome such as SLE, with accompanying serositis. Serologic testing for autoantibodies would be recommended. Peritoneal biopsies, if obtained, may have demonstrated chronic, inflammatory infiltrate (nonspecific) or leukocytoclastic vasculitis (strongly supportive).

ANA enzyme immunoassay was >12 U (normal 1.0 U). Extractable nuclear antigens revealed positive autoantibodies for anti‐SSA, anti‐SSB, and anti‐ribosomal P. Moreover, double‐stranded DNA IgG antibody was 120 IU/mL (normal <30 IU/mL) and C3, C4, and total complement levels were low.

The clinical data support a diagnosis of SLE with serositis. Treatment of the underlying connective tissue disease will typically result in resolution of the ascites; diuretic therapy is generally ineffective.

In consultation with rheumatology and gastroenterology specialists, the diagnosis of SLE was made based on criteria of serositis, persistent leukopenia, arthritis, renal disease (proteinuria), positive ANA, elevated ds‐DNA antibodies, and hypocomplementemia. MR imaging of the abdominal vasculature demonstrated no evidence of vasculitis. The patient was given intravenous methylprednisolone 1 g daily for 3 days followed by high‐dose oral corticosteroids with a gradual taper. He was also started on mycophenolate mofetil as a steroid‐sparing medication (which was later changed to leflunomide due to persistent leukopenia) and hydroxychloroquine. His isolated positive Schistosoma IgG antibody in the absence of other findings was consistent with past exposure or infection. The infectious disease specialist felt there was no evidence of active schistosomiasis, but recommended treatment with a single dose of praziquantel due to the potential benefit with low risk of side effects. The patient had ongoing improvement following dismissal. He had 1 additional paracentesis of 4.1 L, 10 days after his hospitalization, and his ascites and proteinuria resolved. At the 5‐year follow‐up visit, there had been no recurrence of abdominal ascites or abdominal pain. He remains on low‐dose prednisone at 5 mg daily, leflunomide, and hydroxychloroquine.

COMMENTARY

This patient had recurrent ascites with 29.6 L removed over the 4 months prior to admission and an additional 3.4 L during his hospitalization. His outpatient providers initially considered a portal hypertensive etiology of his ascites due to his history of HBV and prior alcohol use. They also appropriately investigated for a possible infectious process. They next directed their evaluation toward the liver biopsy findings, which raised concern for a vascular inflow abnormality. However, the evaluation could have been performed more rapidly and far more cost‐efficiently had a diagnostic paracentesis with calculation of the SAAG been performed early in the evaluation.

The SAAG, which was first described in 1983 by Par and colleagues, is a parameter reflecting the oncotic pressure gradient between the vascular bed and the interstitial splanchnic or ascitic fluid. [1] In the classic study by Runyon and colleagues, a SAAG difference of 1.1 g/dL correctly differentiated causes of ascites due to portal hypertension from those that were not due to portal hypertension 96.7% of the time. [2] Conditions such as nephrotic syndrome, peritoneal carcinomatosis, and serositis (lupus peritonitis) can cause ascites in patients without portal hypertension.

Serositis in the form of pleuritis and/or pericarditis is a common feature of SLE, and ascites has been described in 8% to 11% of SLE patients.[3] However, massive ascites due to lupus peritonitis as a presenting symptom is rare.[4] More common causes of ascites in the setting of SLE include nephrotic syndrome, heart failure, protein‐losing enteropathy, constrictive pericarditis, Budd‐Chiari syndrome, indolent infections such as tuberculosis, and chylous ascites.[5, 6, 7] Of note, lupus peritonitis may be chronic or acute. Chronic ascites develops insidiously with few manifestations of active lupus and may be painless, whereas ascites from acute lupus peritonitis typically develops rapidly and presents with acute abdominal pain and other signs of increased lupus activity.[3, 5, 6, 8, 9]

Ascites from lupus peritonitis may be due to marked serosal exudative accumulation with reduced absorptive capacity in the peritoneum.[3, 4, 10] Other possible causes include peritoneal inflammation from deposition of immune complexes or vasculitis of peritoneal vessels and visceral serous membranes.[4, 9, 11] Although subserosal and submucosal vasculitis have been found in acute ascites, chronic ascites may be related to scarring from vasculitis and serosal inflammation leading to poor venous and lymph drainage.[9] Ascitic fluid characteristics from lupus peritonitis include a SAAG <1.1, presence of white blood cells anywhere in a broad range from 10 to 1630/L, and a range of fluid protein from 3.4 to 4.7 mg/dL.[3] Although not tested in this patient, findings of low complement levels, positive ANA, and elevated anti‐DNA antibody in the ascitic fluid would be supportive of lupus peritonitis, but not specific.[5, 9, 12] Lupus erythematosus cells are occasionally found in the ascitic fluid, but do not rule out other causes of ascites.[9] On retrospective analysis, lupus erythematosus cells were not seen in this patient's pathology specimens.

Treatment of lupus peritonitis and ascites is with high‐dose glucocorticoid therapy, but many patients may need a second immunosuppressant, possibly because of impaired peritoneal circulation from chronic inflammation leading to decreased drug delivery.[13, 14] Chronic ascites may be recalcitrant to systemic glucocorticoids,[3] so a possible alternative therapy is intraperitoneal injection of triamcinolone, which successfully treated massive ascites in a patient who did not respond to oral glucocorticoid treatment.[13] Although ascites may be refractory in some patients, those with chronic lupus peritonitis can generally achieve remission, yet the overall prognosis depends on the presence and severity of multiorgan involvement from SLE. As with any SLE patient, there are also risks of infection from immunosuppression and increased cardiovascular risks.

This patient's evaluation and treatment could have been expedited if he had undergone a paracenteses with determination of the SAAG early in his workup. It is not known why the SAAG was not obtained despite multiple outpatient visits and paracenteses, his history of HBV, and prior alcohol use. This may have been simply an unfortunate oversight. Alternatively, it may have been that his outpatient providers focused on tantalizing clues such as his country of origin, which led to concern for schistosomiasis, and the biopsy findings suggestive of a vascular inflow abnormality that led to further extensive testing. In so doing, the clinicians committed several diagnostic errors, including multiple alternatives bias, anchoring, and confirmation bias.[15] As a result, the patient accrued excess charges of $64,000 from multiple tests, laparoscopic surgery, and 2 hospitalizations. This case highlights how cognitive errors introduce costly variability into patient care, especially when a simple and accurate test is at the beginning of the decision tree.

CLINICAL TEACHING POINTS

1. Diagnostic paracentesis, with calculation of the serum‐ascites albumin gradient, should be the first test in the workup for ascites and can distinguish portal hypertensive causes from nonportal hypertensive causes.
2. Ascites related to SLE can be acute or chronic and caused by bowel infarction, perforation, pancreatitis, mesenteric vasculitis, nephrotic syndrome, heart failure, protein‐losing enteropathy, constrictive pericarditis, lupus peritonitis, Budd‐Chiari syndrome, or serositis (lupus peritonitis).
3. Ascites caused by lupus peritonitis is rare. Once treated, management should be directed toward keeping the SLE in remission.

A 40‐year‐old Sudanese man was admitted due to worsening abdominal pain with recurrent ascites. He had a history of hepatitis B (HBV) infection and diabetes. He previously drank 3 beers per day on the weekends, but he had not consumed alcohol in over a year. He was born in Sudan but lived in Egypt most of his adult life; he immigrated to the United States 6 years previously. He was hospitalized out of state 9 months ago for a swollen abdomen and underwent an exploratory laparotomy that reportedly was unremarkable except for ascites.

Portal hypertension due to liver disease is the most common cause of ascites. This patient has a known risk factor for liver disease (history of HBV infection). Although his reported alcohol consumption is low, there is a synergistic effect on liver injury in the setting of chronic hepatitis. Abdominal pain in the setting of ascites needs to be urgently evaluated to exclude spontaneous bacterial peritonitis (SBP). Also, because chronic HBV infection is the major risk factor for hepatocellular carcinoma in the world, malignant ascites is in the differential. Hepatic vascular thrombosis and tuberculous peritonitis (given the patient's country of origin and travel history) also should be considered. The most appropriate initial test would be a diagnostic paracentesis to support or exclude the presence of SBP and direct the evaluation toward liver disease or other less‐common causes of ascites.

The patient was seen as an outpatient 5 months prior to admission with transient fever and joint pains. Laboratory studies at that visit were notable for a serum albumin of 3.2 g/dL (normal 3.55), 2.4 g of predicted 24‐hour protein on urinalysis (normal <30 mg per 24 hours), creatinine of 0.5 mg/dL (normal 0.81.3), and a positive hepatitis B surface antibody. The working diagnosis was a nonspecific viral syndrome and his symptoms resolved without treatment. One month later, he developed ascites and mild lower extremity edema. Additional laboratory studies at that time showed a normocytic anemia with hemoglobin 11.7 g/dL (normal 13.517.5) and leukopenia with white blood cell count of 2.4 10⁹/L (normal 3.510.5), neutrophil count of 1.45 10⁹/L (normal 1.77.0), and lymphocyte count of 0.58 10⁹/L (normal 0.902.90). Transaminases, serum bilirubin, prothrombin time, alpha fetoprotein, and peripheral blood smear were normal. Human immunodeficiency virus antibody screen and QuantiFERON‐TB assay were negative. Hemoglobin A1c was 6.2% (normal 4.06.0). Repeat urinalysis demonstrated 883 mg of predicted 24‐hour protein. Computed tomography (CT) of the abdomen showed a large amount of intra‐abdominal ascites; the liver and spleen were normal, and there were no varices or other evidence of portal hypertension. Echocardiogram was normal except for a small inferior vena cava (IVC) and a mildly increased right ventricular systolic pressure of 32 mm Hg (systolic blood pressure 98 mm Hg). Due to the indeterminate cause for the patient's ascites, referral was made for gastroenterology evaluation with consideration for a paracentesis.

Cirrhotic ascites seems less likely. Postsinusoidal causes of portal hypertension (eg, cardiomyopathy) are also less likely given the absence of suggestive findings on echocardiography. Malignant ascites also appears less probable in the absence of suggestive findings such as mass lesions, lymphadenopathy, or peritoneal carcinomatosis on CT imaging. The suspicion for tuberculous peritonitis is lower with the negative QuantiFERON‐TB test. Hypoalbuminemia, normocytic anemia, leukopenia, and proteinuria all suggest a systemic inflammatory condition (eg, systemic lupus erythematosus [SLE]) with inflammatory serositis causing ascites). Nephrotic syndrome can cause hypoalbuminemia, edema, and ascites, but his total urine protein losses of <3.5 grams per 24 hours are not in keeping with this diagnosis. Other uncommon causes of ascites such as chylous ascites have not yet been excluded. The most appropriate next step remains ascitic fluid analysis.

A paracentesis yielded 7.8 L of clear‐yellow fluid and improvement in his abdominal discomfort. Analysis showed 224 total nucleated cells/L with 2% neutrophils, 57% lymphocytes, and 37% monocytes. Ascites total protein was 3.8 g/dL and glucose was 55 mg/dL. Gram stain and culture were negative, and cytology was negative for malignancy but showed lymphocytes, plasma cells, monocytes, and reactive mesothelial cells interpreted as consistent with chronic inflammation. The serum‐ascites albumin gradient (SAAG) was not obtained.

With a low leukocyte count and a paucity of neutrophils, this is not SBP. The ascites fluid did not have a chylous appearance. The SAAG, which can distinguish between portal hypertensive and nonportal hypertensive causes for ascites using a cutoff of 1.1 g/dL, was not done. The total protein was high, arguing against cirrhosis. High protein ascites with a high SAAG would suggest a posthepatic source of portal hypertension (eg, Budd‐Chiari syndrome, constrictive pericarditis). High protein ascites with a low SAAG would suggest an inflammatory or malignant source of ascites. The relative lymphocytosis in the ascites fluid suggests an inflammatory process, but is a nonspecific finding. The negative cytology does not completely exclude a malignancy, but given the absence of findings on the CT, malignant ascites is less likely.

Three months before admission, the patient underwent a repeat large‐volume paracentesis and a liver biopsy. The biopsy showed ectopic portal vein branches consistent with hepatoportal sclerosis, but no actual sclerosis was identified. The pathologist concluded that the findings suggested noncirrhotic portal hypertension due to a vascular in‐flow abnormality. Abdominal ultrasound with Doppler was unremarkable other than slightly increased echogenicity of the liver. Magnetic resonance (MR) angiogram showed narrowing of the intra‐abdominal IVC at the level of the diaphragm. Because of concern that hepatic congestion from high pressures in the narrowed IVC was leading to poor vascular inflow as suggested by the biopsy findings, an inferior vena cavagram was performed. This study was normal, although no transhepatic pressure measurements were obtained. Three stool specimens and 2 urine specimens were negative for parasites. The patient required repeat large‐volume paracenteses monthly. SBP was again ruled out, but no other diagnostic labs were obtained. He had anorexia with poor oral intake each time his abdomen became distended.

The patient was started on furosemide 1 month prior to admission to the hospital but had only a slight improvement in the ascites. His other medications included insulin, tamsulosin, and hydrocodone‐acetaminophen. Five days prior to admission, he underwent a diagnostic laparoscopy, which showed only ascites and small adhesions to the anterior abdominal wall. There was no visual evidence of malignancy, and the surgeon commented that the liver was normal. No additional biopsies were obtained.

The liver biopsy findings could be seen in noncirrhotic portal hypertension, although this diagnosis would be unlikely without splenomegaly, varices, or other signs of portal hypertension. However, 2 possible etiologies for noncirrhotic portal hypertension in this patient would be hepatic congestion from the narrowed IVC (although the normal IVC study argues against this) and hepatic schistosomiasis. Schistosomiasis is an important cause of noncirrhotic portal hypertension in endemic areas like this patient's country of origin, but the negative stool and urine studies, combined with the lack of granulomas or fibrosis seen on biopsy, make this condition unlikely.

Systemic amyloidosis (primary or secondary) could also be a cause of ascites and could present with multiorgan involvement (diarrhea and nephrotic syndrome). Amyloid deposits would have probably been seen in the liver biopsy, if present, but may not have been apparent unless specific stains (Congo red) were performed.

Evaluation for systemic, inflammatory autoimmune processes is indicated. Serum autoantibodies (anti‐nuclear antibody [ANA] and extractable nuclear antigens), and a serum and 24‐hour urine protein electrophoresis would be appropriate diagnostic tests. Peritoneal biopsies would have been helpful to assess for serosal diseases.

The patient subsequently developed acute right‐sided abdominal pain requiring urgent evaluation and admission to the hospital. He was initially assessed by a general surgeon, who found no evidence of postoperative complications. His temperature was 36.7C, blood pressure 105/64, heart rate 82, respiratory rate 16, and oxygen saturation 97% on room air. He appeared chronically ill, but he was in no distress and he had a normal mental status. Cardiac exam was normal except for mild jugular venous distension. He had mild bibasilar lung crackles. His abdomen was distended with superficial abdominal tenderness and a fluid wave, but he had normal bowel sounds and no peritoneal signs. He had mild scrotal edema but no peripheral edema. Joint exam did not suggest synovitis and there were no rashes or oral ulcers. Lactate was 0.9 mmol/L (normal 0.62.3), albumin was 2.6 g/dL, and prealbumin was 9 mg/dL (normal 1938). Erythrocyte sedimentation rate and C‐reactive protein were 46 mm/hour (normal <22) and 33.1 mg/L (normal 8), respectively. He had a normocytic anemia and leukopenia. Liver tests and routine chemistries were normal. Serum protein electrophoresis indicated no monoclonal protein. Complete 24‐hour urine collection showed 1.2 g of protein (normal <102 mg). Paracentesis of 3.4 L demonstrated 227 total nucleated cells/L with 2% neutrophils. Following the fluid removal, he had improvement in his pain, which he felt was related to the ascites rather than the recent surgery. Ascites total protein was 3.9 g/dL and ascites albumin was 1.7 g/dL. Ascites culture was negative for infection. Serum Schistosoma immunoglobulin G (IgG) antibody was positive at 3.53 (normal <1.00).

Further history revealed prior episodes of polyarticular joint pain and swelling in his hands and knees 5 years before admission. At that time, he reported a diffuse, pruritic, papular body rash. In addition, he noticed that his fingertips and toes turned white with cold exposure.

Importantly, surgical and infectious complications have been excluded. High protein ascites with a low SAAG of 0.9 suggests an inflammatory source of ascites. The follow‐up clinical data (arthritis, normocytic anemia, leukopenia, rash, Raynaud's phenomenon) suggest a systemic inflammatory syndrome such as SLE, with accompanying serositis. Serologic testing for autoantibodies would be recommended. Peritoneal biopsies, if obtained, may have demonstrated chronic, inflammatory infiltrate (nonspecific) or leukocytoclastic vasculitis (strongly supportive).

ANA enzyme immunoassay was >12 U (normal 1.0 U). Extractable nuclear antigens revealed positive autoantibodies for anti‐SSA, anti‐SSB, and anti‐ribosomal P. Moreover, double‐stranded DNA IgG antibody was 120 IU/mL (normal <30 IU/mL) and C3, C4, and total complement levels were low.

The clinical data support a diagnosis of SLE with serositis. Treatment of the underlying connective tissue disease will typically result in resolution of the ascites; diuretic therapy is generally ineffective.

In consultation with rheumatology and gastroenterology specialists, the diagnosis of SLE was made based on criteria of serositis, persistent leukopenia, arthritis, renal disease (proteinuria), positive ANA, elevated ds‐DNA antibodies, and hypocomplementemia. MR imaging of the abdominal vasculature demonstrated no evidence of vasculitis. The patient was given intravenous methylprednisolone 1 g daily for 3 days followed by high‐dose oral corticosteroids with a gradual taper. He was also started on mycophenolate mofetil as a steroid‐sparing medication (which was later changed to leflunomide due to persistent leukopenia) and hydroxychloroquine. His isolated positive Schistosoma IgG antibody in the absence of other findings was consistent with past exposure or infection. The infectious disease specialist felt there was no evidence of active schistosomiasis, but recommended treatment with a single dose of praziquantel due to the potential benefit with low risk of side effects. The patient had ongoing improvement following dismissal. He had 1 additional paracentesis of 4.1 L, 10 days after his hospitalization, and his ascites and proteinuria resolved. At the 5‐year follow‐up visit, there had been no recurrence of abdominal ascites or abdominal pain. He remains on low‐dose prednisone at 5 mg daily, leflunomide, and hydroxychloroquine.

COMMENTARY

This patient had recurrent ascites with 29.6 L removed over the 4 months prior to admission and an additional 3.4 L during his hospitalization. His outpatient providers initially considered a portal hypertensive etiology of his ascites due to his history of HBV and prior alcohol use. They also appropriately investigated for a possible infectious process. They next directed their evaluation toward the liver biopsy findings, which raised concern for a vascular inflow abnormality. However, the evaluation could have been performed more rapidly and far more cost‐efficiently had a diagnostic paracentesis with calculation of the SAAG been performed early in the evaluation.

The SAAG, which was first described in 1983 by Par and colleagues, is a parameter reflecting the oncotic pressure gradient between the vascular bed and the interstitial splanchnic or ascitic fluid. [1] In the classic study by Runyon and colleagues, a SAAG difference of 1.1 g/dL correctly differentiated causes of ascites due to portal hypertension from those that were not due to portal hypertension 96.7% of the time. [2] Conditions such as nephrotic syndrome, peritoneal carcinomatosis, and serositis (lupus peritonitis) can cause ascites in patients without portal hypertension.

Serositis in the form of pleuritis and/or pericarditis is a common feature of SLE, and ascites has been described in 8% to 11% of SLE patients.[3] However, massive ascites due to lupus peritonitis as a presenting symptom is rare.[4] More common causes of ascites in the setting of SLE include nephrotic syndrome, heart failure, protein‐losing enteropathy, constrictive pericarditis, Budd‐Chiari syndrome, indolent infections such as tuberculosis, and chylous ascites.[5, 6, 7] Of note, lupus peritonitis may be chronic or acute. Chronic ascites develops insidiously with few manifestations of active lupus and may be painless, whereas ascites from acute lupus peritonitis typically develops rapidly and presents with acute abdominal pain and other signs of increased lupus activity.[3, 5, 6, 8, 9]

Ascites from lupus peritonitis may be due to marked serosal exudative accumulation with reduced absorptive capacity in the peritoneum.[3, 4, 10] Other possible causes include peritoneal inflammation from deposition of immune complexes or vasculitis of peritoneal vessels and visceral serous membranes.[4, 9, 11] Although subserosal and submucosal vasculitis have been found in acute ascites, chronic ascites may be related to scarring from vasculitis and serosal inflammation leading to poor venous and lymph drainage.[9] Ascitic fluid characteristics from lupus peritonitis include a SAAG <1.1, presence of white blood cells anywhere in a broad range from 10 to 1630/L, and a range of fluid protein from 3.4 to 4.7 mg/dL.[3] Although not tested in this patient, findings of low complement levels, positive ANA, and elevated anti‐DNA antibody in the ascitic fluid would be supportive of lupus peritonitis, but not specific.[5, 9, 12] Lupus erythematosus cells are occasionally found in the ascitic fluid, but do not rule out other causes of ascites.[9] On retrospective analysis, lupus erythematosus cells were not seen in this patient's pathology specimens.

Treatment of lupus peritonitis and ascites is with high‐dose glucocorticoid therapy, but many patients may need a second immunosuppressant, possibly because of impaired peritoneal circulation from chronic inflammation leading to decreased drug delivery.[13, 14] Chronic ascites may be recalcitrant to systemic glucocorticoids,[3] so a possible alternative therapy is intraperitoneal injection of triamcinolone, which successfully treated massive ascites in a patient who did not respond to oral glucocorticoid treatment.[13] Although ascites may be refractory in some patients, those with chronic lupus peritonitis can generally achieve remission, yet the overall prognosis depends on the presence and severity of multiorgan involvement from SLE. As with any SLE patient, there are also risks of infection from immunosuppression and increased cardiovascular risks.

This patient's evaluation and treatment could have been expedited if he had undergone a paracenteses with determination of the SAAG early in his workup. It is not known why the SAAG was not obtained despite multiple outpatient visits and paracenteses, his history of HBV, and prior alcohol use. This may have been simply an unfortunate oversight. Alternatively, it may have been that his outpatient providers focused on tantalizing clues such as his country of origin, which led to concern for schistosomiasis, and the biopsy findings suggestive of a vascular inflow abnormality that led to further extensive testing. In so doing, the clinicians committed several diagnostic errors, including multiple alternatives bias, anchoring, and confirmation bias.[15] As a result, the patient accrued excess charges of $64,000 from multiple tests, laparoscopic surgery, and 2 hospitalizations. This case highlights how cognitive errors introduce costly variability into patient care, especially when a simple and accurate test is at the beginning of the decision tree.

CLINICAL TEACHING POINTS

1. Diagnostic paracentesis, with calculation of the serum‐ascites albumin gradient, should be the first test in the workup for ascites and can distinguish portal hypertensive causes from nonportal hypertensive causes.
2. Ascites related to SLE can be acute or chronic and caused by bowel infarction, perforation, pancreatitis, mesenteric vasculitis, nephrotic syndrome, heart failure, protein‐losing enteropathy, constrictive pericarditis, lupus peritonitis, Budd‐Chiari syndrome, or serositis (lupus peritonitis).
3. Ascites caused by lupus peritonitis is rare. Once treated, management should be directed toward keeping the SLE in remission.

Acute change in neurologic status in a hospitalized patient is an emergency requiring timely coordinated evaluation. To address this need, many hospitals have created a mechanism for in‐hospital stroke alerts utilizing generalized rapid response teams or specialized stroke teams.[1, 2, 3] The common purpose is to quickly diagnose new ischemic stroke within the time window for thrombolytic therapy.

Even when acute change in neurologic status is not due to brain ischemia, it may represent a new metabolic disturbance or reflect developing serious systemic illness. Sepsis, hypoglycemia, cardiac arrhythmia, respiratory failure, severe electrolyte disturbances, seizures, or delirium may first manifest as a change in neurologic status.

Prior research on stroke alerts has largely focused on patients who present from the community to the emergency department (ED).[4, 5, 6, 7, 8] Patients who develop acute neurologic symptoms during hospitalization have different risk factors and exposures compared to patients in the community.[9] This study represents the experience of a multistate quality improvement initiative for in‐hospital stroke. We characterize etiologies for symptoms triggering in‐hospital stroke alerts and thrombolytic treatment for in‐hospital strokes.

PATIENTS AND METHODS

The National Stroke Association's (NSA) initiative, Improving In‐Hospital Stroke Response: A Team‐based Quality Improvement Program, included data collection for all in‐hospital stroke alerts over a 12‐month period.[10] Six Joint Commission certified primary stroke centers from Michigan, South Carolina, Pennsylvania, Colorado, Washington, and North Carolina completed the 1‐year quality improvement initiative. One additional site withdrew from the program after the first quarter and was not included in this analysis. Sites prospectively reported deidentified patient‐level data on all adult in‐hospital stroke alerts from July 2010 to June 2011 to the NSA. At all sites, any provider could activate the in‐hospital stroke response system. Stroke alerts were evaluated by a rapid response team with stroke training. The providers on the stroke rapid response team varied between sites. A nurse with stroke training was 1 of the first responders on the stroke response team at all sites.

The NSA in‐hospital stroke‐alert criteria included the following symptoms occurring in the last 24‐hours, even if they resolved: (1) sudden numbness or weakness of the face, arm or leg, especially on 1 side of the body; (2) sudden confusion, trouble speaking or understanding; (3) sudden trouble seeing in 1 or both eyes; (4) sudden trouble walking, dizziness, loss of balance or coordination; and (5) sudden, severe headache with no known cause. Hospitals reported location, service, age, sex, race, symptoms triggering the stroke alert, free text entry of final clinical diagnosis following the completion of stroke alert evaluation, treatment with intravenous or intra‐arterial/mechanical thrombolysis, and any contraindications to intravenous thrombolysis. We categorized stroke mimics using the responses in the final diagnosis field after the data collection period was complete. Strokes were categorized as ischemic stroke, transient ischemic attack (TIA), or intracranial hemorrhage (intraparenchymal, intraventricular, epidural, subdural, or subarachnoid). Stroke mimics were subdivided according to the categories in Table 1. Lack of certainty in the final diagnosis was handled by creating a category of possible TIA, which includes alternative diagnosis versus TIA or the qualifier possible before TIA. Patients with final diagnoses unable to be determined were classified as stroke mimics. Institutional review board exemption was obtained for the deidentified prospective data registry of this quality‐improvement program.

RESULTS

During the 12‐month data collection period, 393 in‐hospital stroke alerts were reported to the NSA. Hospitals reported an average of 65.5 in‐hospital stroke alerts (range, 27156; standard deviation 46.8) (Table 2). Median age was 70 years (range, 18 to >89 years, interquartile range [IQR], 6280 years). Of the stoke alert patients, 52.8% were female, 81.7% were white, 12.7% were black, 2.9% were Hispanic, and 2.7% were other or were unable to be determined. The most common primary services were medicine/hospitalist (36.4%), cardiology (19.5%), cardiothoracic/vascular surgery (13%), and orthopedic surgery (8.6%).

Of the stroke alert patients, 167 (42.5%) were found to have ischemic stroke, 27 (6.9%) TIA, 11 (2.8%) intracranial hemorrhage, and 7 (1.8%) had TIA possible or considered along with a stroke mimic in the final diagnosis. The stroke mimic rate was 46.1%, with a confidence range of 42.0% to 47.8% depending on the true pathologic cause of the alerts in the categories possible TIA and final diagnosis uncertain. Participating hospitals had an alarm rate for stroke mimics ranging from 28.0% to 66.7% (median, 45.8%; IQR, 32.9%49.7%) (Table 2). The most common stroke mimics were seizure, hypotension, and delirium (Table 1). Data were available on symptoms that triggered the alert in 373 (94.9%) of cases. Eighteen alerts (4.8%) were for symptoms clearly not included in the NSA stroke alert criteria. The final diagnosis was acute ischemic stroke/TIA or intracranial hemorrhage in 4 of these 18 (22.2%) nonconforming alerts. If alerts called for a decrease in consciousness were also considered nonconforming, then 67 alerts (18.0%) could be categorized as nonconforming. However, 24 of these 67 alerts (35.8%) had a final diagnosis of acute ischemic stroke/TIA or intracranial hemorrhage.

For 194 patients with a final diagnosis of ischemic stroke or TIA, intravenous thrombolysis alone was used for 16 in‐hospital stroke patients (8.2%), 20 received intra‐arterial/mechanical thrombolysis alone (10.3%), and 2 patients received both (1%) (Table 3). No patient with a stroke mimic received thrombolysis.

DISCUSSION

Given the protean manifestations of brain ischemia, and significant symptom overlap with many mimics, stroke alert criteria casts a wide net in order not to miss or delay evaluation and treatment of true brain ischemia. Time is critical given the association of improved outcomes with more rapid delivery of treatment.[11] The inevitable consequence of the combination of time pressure and clinical uncertainty based solely on physical exam will be alerts due to stroke mimics. Our analysis reveals many of these alternative diagnoses also require urgent evaluation and treatment.

Prior research has found a large proportion of in‐hospital stroke alerts are not for cerebrovascular events.[1, 4, 12] We observed an average of 46.1% of in‐hospital stroke alerts were due to mimics. This rate is substantially higher than described in studies of stroke mimics in the ED.[7, 13, 14] The largest analysis over a 10‐year period from 2 hospitals in Washington found a 30% stroke mimic rate and concluded that in‐hospital location for symptom onset was a statistically significant predictor of being a mimic rather than a cerebrovascular event.[4] One single‐center trial in North Carolina found markedly higher mimic rates for in‐hospital stroke alerts (73%) versus ED stroke alerts (49%).[12] Assessment of neurologic symptoms is challenging in patients already hospitalized for acute medical conditions. The interaction of systemic illness, medications, and surgery seen in the hospital setting may make it more difficult to distinguish between cerebrovascular events and their many mimics.

Interpretation of NSA criteria for calling a stroke code likely varied within and between sites, and inter‐rater reliability of physical signs was not assessed, which is a limitation of the data. Observed rates of stroke for alerts that did not conform to the NSA criteria suggest that clinical judgment remains valuable. Final diagnoses were assigned by the stroke programs, and reliability of this assessment was not evaluated. Sites were not asked to use a specific categorization scheme to group final diagnoses. This analysis was limited to stroke centers with existing infrastructure to respond to stroke alerts and participated in an explicit quality‐improvement initiative on in‐hospital stroke response. Mimic and thrombolysis treatment rates may be different for hospitals without this stroke expertise.

Clinical uncertainty as to final diagnosis was addressed with the inclusion of confidence intervals accounting for potential misdiagnosis of the events in the categories of possible TIA or in the cases where the final diagnosis was unknown. Other studies have categorized TIA versus an alternative diagnosis as stroke mimic, and so our methodology is expected to yield a conservative estimate of the stroke mimic rate. Delirium is often a multifactorial phenomenon, so there may be an element of overlap between this category and other more specific mimic etiologies such as infection, hypotension, metabolic, or medication effect.

This initiative did not have the ability to assess the false negative rate of stroke team activation (failure to identify stroke symptoms in time for acute evaluation). It is not possible to calculate the sensitivity of stroke alerts in each center or conclude the optimal rate of false alarms. The finding of inter‐institutional variability in stroke alerts due to true brain ischemia could be explained by differences in staff education, systematic differences in the patient populations cared for among hospitals, or variation in institutional acceptance of having activated the stroke response team for cases with lower pretest probability of stroke. Sensitivity of alert criteria is more important than specificity, given the consequences of missing a potentially treatable emergent condition.

In conclusion, in this multi‐institution analysis of in‐hospital stroke alerts, a substantial proportion of in‐hospital strokes received thrombolytic therapy. Almost half of stroke alerts will not be for stroke or TIA. For many patients in our study, a change in neurologic status represented a harbinger of a change in general medical condition (hemorrhage, hypotension, hypoglycemia, or respiratory failure). Rapid response systems used for stroke in the hospital need to be trained and prepared to respond to a variety of acute medical conditions that extend beyond ischemic stroke.

Acute change in neurologic status in a hospitalized patient is an emergency requiring timely coordinated evaluation. To address this need, many hospitals have created a mechanism for in‐hospital stroke alerts utilizing generalized rapid response teams or specialized stroke teams.[1, 2, 3] The common purpose is to quickly diagnose new ischemic stroke within the time window for thrombolytic therapy.

Even when acute change in neurologic status is not due to brain ischemia, it may represent a new metabolic disturbance or reflect developing serious systemic illness. Sepsis, hypoglycemia, cardiac arrhythmia, respiratory failure, severe electrolyte disturbances, seizures, or delirium may first manifest as a change in neurologic status.

Prior research on stroke alerts has largely focused on patients who present from the community to the emergency department (ED).[4, 5, 6, 7, 8] Patients who develop acute neurologic symptoms during hospitalization have different risk factors and exposures compared to patients in the community.[9] This study represents the experience of a multistate quality improvement initiative for in‐hospital stroke. We characterize etiologies for symptoms triggering in‐hospital stroke alerts and thrombolytic treatment for in‐hospital strokes.

PATIENTS AND METHODS

The National Stroke Association's (NSA) initiative, Improving In‐Hospital Stroke Response: A Team‐based Quality Improvement Program, included data collection for all in‐hospital stroke alerts over a 12‐month period.[10] Six Joint Commission certified primary stroke centers from Michigan, South Carolina, Pennsylvania, Colorado, Washington, and North Carolina completed the 1‐year quality improvement initiative. One additional site withdrew from the program after the first quarter and was not included in this analysis. Sites prospectively reported deidentified patient‐level data on all adult in‐hospital stroke alerts from July 2010 to June 2011 to the NSA. At all sites, any provider could activate the in‐hospital stroke response system. Stroke alerts were evaluated by a rapid response team with stroke training. The providers on the stroke rapid response team varied between sites. A nurse with stroke training was 1 of the first responders on the stroke response team at all sites.

The NSA in‐hospital stroke‐alert criteria included the following symptoms occurring in the last 24‐hours, even if they resolved: (1) sudden numbness or weakness of the face, arm or leg, especially on 1 side of the body; (2) sudden confusion, trouble speaking or understanding; (3) sudden trouble seeing in 1 or both eyes; (4) sudden trouble walking, dizziness, loss of balance or coordination; and (5) sudden, severe headache with no known cause. Hospitals reported location, service, age, sex, race, symptoms triggering the stroke alert, free text entry of final clinical diagnosis following the completion of stroke alert evaluation, treatment with intravenous or intra‐arterial/mechanical thrombolysis, and any contraindications to intravenous thrombolysis. We categorized stroke mimics using the responses in the final diagnosis field after the data collection period was complete. Strokes were categorized as ischemic stroke, transient ischemic attack (TIA), or intracranial hemorrhage (intraparenchymal, intraventricular, epidural, subdural, or subarachnoid). Stroke mimics were subdivided according to the categories in Table 1. Lack of certainty in the final diagnosis was handled by creating a category of possible TIA, which includes alternative diagnosis versus TIA or the qualifier possible before TIA. Patients with final diagnoses unable to be determined were classified as stroke mimics. Institutional review board exemption was obtained for the deidentified prospective data registry of this quality‐improvement program.

RESULTS

During the 12‐month data collection period, 393 in‐hospital stroke alerts were reported to the NSA. Hospitals reported an average of 65.5 in‐hospital stroke alerts (range, 27156; standard deviation 46.8) (Table 2). Median age was 70 years (range, 18 to >89 years, interquartile range [IQR], 6280 years). Of the stoke alert patients, 52.8% were female, 81.7% were white, 12.7% were black, 2.9% were Hispanic, and 2.7% were other or were unable to be determined. The most common primary services were medicine/hospitalist (36.4%), cardiology (19.5%), cardiothoracic/vascular surgery (13%), and orthopedic surgery (8.6%).

Of the stroke alert patients, 167 (42.5%) were found to have ischemic stroke, 27 (6.9%) TIA, 11 (2.8%) intracranial hemorrhage, and 7 (1.8%) had TIA possible or considered along with a stroke mimic in the final diagnosis. The stroke mimic rate was 46.1%, with a confidence range of 42.0% to 47.8% depending on the true pathologic cause of the alerts in the categories possible TIA and final diagnosis uncertain. Participating hospitals had an alarm rate for stroke mimics ranging from 28.0% to 66.7% (median, 45.8%; IQR, 32.9%49.7%) (Table 2). The most common stroke mimics were seizure, hypotension, and delirium (Table 1). Data were available on symptoms that triggered the alert in 373 (94.9%) of cases. Eighteen alerts (4.8%) were for symptoms clearly not included in the NSA stroke alert criteria. The final diagnosis was acute ischemic stroke/TIA or intracranial hemorrhage in 4 of these 18 (22.2%) nonconforming alerts. If alerts called for a decrease in consciousness were also considered nonconforming, then 67 alerts (18.0%) could be categorized as nonconforming. However, 24 of these 67 alerts (35.8%) had a final diagnosis of acute ischemic stroke/TIA or intracranial hemorrhage.

For 194 patients with a final diagnosis of ischemic stroke or TIA, intravenous thrombolysis alone was used for 16 in‐hospital stroke patients (8.2%), 20 received intra‐arterial/mechanical thrombolysis alone (10.3%), and 2 patients received both (1%) (Table 3). No patient with a stroke mimic received thrombolysis.

DISCUSSION

Given the protean manifestations of brain ischemia, and significant symptom overlap with many mimics, stroke alert criteria casts a wide net in order not to miss or delay evaluation and treatment of true brain ischemia. Time is critical given the association of improved outcomes with more rapid delivery of treatment.[11] The inevitable consequence of the combination of time pressure and clinical uncertainty based solely on physical exam will be alerts due to stroke mimics. Our analysis reveals many of these alternative diagnoses also require urgent evaluation and treatment.

Prior research has found a large proportion of in‐hospital stroke alerts are not for cerebrovascular events.[1, 4, 12] We observed an average of 46.1% of in‐hospital stroke alerts were due to mimics. This rate is substantially higher than described in studies of stroke mimics in the ED.[7, 13, 14] The largest analysis over a 10‐year period from 2 hospitals in Washington found a 30% stroke mimic rate and concluded that in‐hospital location for symptom onset was a statistically significant predictor of being a mimic rather than a cerebrovascular event.[4] One single‐center trial in North Carolina found markedly higher mimic rates for in‐hospital stroke alerts (73%) versus ED stroke alerts (49%).[12] Assessment of neurologic symptoms is challenging in patients already hospitalized for acute medical conditions. The interaction of systemic illness, medications, and surgery seen in the hospital setting may make it more difficult to distinguish between cerebrovascular events and their many mimics.

Interpretation of NSA criteria for calling a stroke code likely varied within and between sites, and inter‐rater reliability of physical signs was not assessed, which is a limitation of the data. Observed rates of stroke for alerts that did not conform to the NSA criteria suggest that clinical judgment remains valuable. Final diagnoses were assigned by the stroke programs, and reliability of this assessment was not evaluated. Sites were not asked to use a specific categorization scheme to group final diagnoses. This analysis was limited to stroke centers with existing infrastructure to respond to stroke alerts and participated in an explicit quality‐improvement initiative on in‐hospital stroke response. Mimic and thrombolysis treatment rates may be different for hospitals without this stroke expertise.

Clinical uncertainty as to final diagnosis was addressed with the inclusion of confidence intervals accounting for potential misdiagnosis of the events in the categories of possible TIA or in the cases where the final diagnosis was unknown. Other studies have categorized TIA versus an alternative diagnosis as stroke mimic, and so our methodology is expected to yield a conservative estimate of the stroke mimic rate. Delirium is often a multifactorial phenomenon, so there may be an element of overlap between this category and other more specific mimic etiologies such as infection, hypotension, metabolic, or medication effect.

This initiative did not have the ability to assess the false negative rate of stroke team activation (failure to identify stroke symptoms in time for acute evaluation). It is not possible to calculate the sensitivity of stroke alerts in each center or conclude the optimal rate of false alarms. The finding of inter‐institutional variability in stroke alerts due to true brain ischemia could be explained by differences in staff education, systematic differences in the patient populations cared for among hospitals, or variation in institutional acceptance of having activated the stroke response team for cases with lower pretest probability of stroke. Sensitivity of alert criteria is more important than specificity, given the consequences of missing a potentially treatable emergent condition.

In conclusion, in this multi‐institution analysis of in‐hospital stroke alerts, a substantial proportion of in‐hospital strokes received thrombolytic therapy. Almost half of stroke alerts will not be for stroke or TIA. For many patients in our study, a change in neurologic status represented a harbinger of a change in general medical condition (hemorrhage, hypotension, hypoglycemia, or respiratory failure). Rapid response systems used for stroke in the hospital need to be trained and prepared to respond to a variety of acute medical conditions that extend beyond ischemic stroke.

Acute change in neurologic status in a hospitalized patient is an emergency requiring timely coordinated evaluation. To address this need, many hospitals have created a mechanism for in‐hospital stroke alerts utilizing generalized rapid response teams or specialized stroke teams.[1, 2, 3] The common purpose is to quickly diagnose new ischemic stroke within the time window for thrombolytic therapy.

Even when acute change in neurologic status is not due to brain ischemia, it may represent a new metabolic disturbance or reflect developing serious systemic illness. Sepsis, hypoglycemia, cardiac arrhythmia, respiratory failure, severe electrolyte disturbances, seizures, or delirium may first manifest as a change in neurologic status.

Prior research on stroke alerts has largely focused on patients who present from the community to the emergency department (ED).[4, 5, 6, 7, 8] Patients who develop acute neurologic symptoms during hospitalization have different risk factors and exposures compared to patients in the community.[9] This study represents the experience of a multistate quality improvement initiative for in‐hospital stroke. We characterize etiologies for symptoms triggering in‐hospital stroke alerts and thrombolytic treatment for in‐hospital strokes.

PATIENTS AND METHODS

The National Stroke Association's (NSA) initiative, Improving In‐Hospital Stroke Response: A Team‐based Quality Improvement Program, included data collection for all in‐hospital stroke alerts over a 12‐month period.[10] Six Joint Commission certified primary stroke centers from Michigan, South Carolina, Pennsylvania, Colorado, Washington, and North Carolina completed the 1‐year quality improvement initiative. One additional site withdrew from the program after the first quarter and was not included in this analysis. Sites prospectively reported deidentified patient‐level data on all adult in‐hospital stroke alerts from July 2010 to June 2011 to the NSA. At all sites, any provider could activate the in‐hospital stroke response system. Stroke alerts were evaluated by a rapid response team with stroke training. The providers on the stroke rapid response team varied between sites. A nurse with stroke training was 1 of the first responders on the stroke response team at all sites.

The NSA in‐hospital stroke‐alert criteria included the following symptoms occurring in the last 24‐hours, even if they resolved: (1) sudden numbness or weakness of the face, arm or leg, especially on 1 side of the body; (2) sudden confusion, trouble speaking or understanding; (3) sudden trouble seeing in 1 or both eyes; (4) sudden trouble walking, dizziness, loss of balance or coordination; and (5) sudden, severe headache with no known cause. Hospitals reported location, service, age, sex, race, symptoms triggering the stroke alert, free text entry of final clinical diagnosis following the completion of stroke alert evaluation, treatment with intravenous or intra‐arterial/mechanical thrombolysis, and any contraindications to intravenous thrombolysis. We categorized stroke mimics using the responses in the final diagnosis field after the data collection period was complete. Strokes were categorized as ischemic stroke, transient ischemic attack (TIA), or intracranial hemorrhage (intraparenchymal, intraventricular, epidural, subdural, or subarachnoid). Stroke mimics were subdivided according to the categories in Table 1. Lack of certainty in the final diagnosis was handled by creating a category of possible TIA, which includes alternative diagnosis versus TIA or the qualifier possible before TIA. Patients with final diagnoses unable to be determined were classified as stroke mimics. Institutional review board exemption was obtained for the deidentified prospective data registry of this quality‐improvement program.

RESULTS

During the 12‐month data collection period, 393 in‐hospital stroke alerts were reported to the NSA. Hospitals reported an average of 65.5 in‐hospital stroke alerts (range, 27156; standard deviation 46.8) (Table 2). Median age was 70 years (range, 18 to >89 years, interquartile range [IQR], 6280 years). Of the stoke alert patients, 52.8% were female, 81.7% were white, 12.7% were black, 2.9% were Hispanic, and 2.7% were other or were unable to be determined. The most common primary services were medicine/hospitalist (36.4%), cardiology (19.5%), cardiothoracic/vascular surgery (13%), and orthopedic surgery (8.6%).

Of the stroke alert patients, 167 (42.5%) were found to have ischemic stroke, 27 (6.9%) TIA, 11 (2.8%) intracranial hemorrhage, and 7 (1.8%) had TIA possible or considered along with a stroke mimic in the final diagnosis. The stroke mimic rate was 46.1%, with a confidence range of 42.0% to 47.8% depending on the true pathologic cause of the alerts in the categories possible TIA and final diagnosis uncertain. Participating hospitals had an alarm rate for stroke mimics ranging from 28.0% to 66.7% (median, 45.8%; IQR, 32.9%49.7%) (Table 2). The most common stroke mimics were seizure, hypotension, and delirium (Table 1). Data were available on symptoms that triggered the alert in 373 (94.9%) of cases. Eighteen alerts (4.8%) were for symptoms clearly not included in the NSA stroke alert criteria. The final diagnosis was acute ischemic stroke/TIA or intracranial hemorrhage in 4 of these 18 (22.2%) nonconforming alerts. If alerts called for a decrease in consciousness were also considered nonconforming, then 67 alerts (18.0%) could be categorized as nonconforming. However, 24 of these 67 alerts (35.8%) had a final diagnosis of acute ischemic stroke/TIA or intracranial hemorrhage.

For 194 patients with a final diagnosis of ischemic stroke or TIA, intravenous thrombolysis alone was used for 16 in‐hospital stroke patients (8.2%), 20 received intra‐arterial/mechanical thrombolysis alone (10.3%), and 2 patients received both (1%) (Table 3). No patient with a stroke mimic received thrombolysis.

DISCUSSION

Given the protean manifestations of brain ischemia, and significant symptom overlap with many mimics, stroke alert criteria casts a wide net in order not to miss or delay evaluation and treatment of true brain ischemia. Time is critical given the association of improved outcomes with more rapid delivery of treatment.[11] The inevitable consequence of the combination of time pressure and clinical uncertainty based solely on physical exam will be alerts due to stroke mimics. Our analysis reveals many of these alternative diagnoses also require urgent evaluation and treatment.

Prior research has found a large proportion of in‐hospital stroke alerts are not for cerebrovascular events.[1, 4, 12] We observed an average of 46.1% of in‐hospital stroke alerts were due to mimics. This rate is substantially higher than described in studies of stroke mimics in the ED.[7, 13, 14] The largest analysis over a 10‐year period from 2 hospitals in Washington found a 30% stroke mimic rate and concluded that in‐hospital location for symptom onset was a statistically significant predictor of being a mimic rather than a cerebrovascular event.[4] One single‐center trial in North Carolina found markedly higher mimic rates for in‐hospital stroke alerts (73%) versus ED stroke alerts (49%).[12] Assessment of neurologic symptoms is challenging in patients already hospitalized for acute medical conditions. The interaction of systemic illness, medications, and surgery seen in the hospital setting may make it more difficult to distinguish between cerebrovascular events and their many mimics.

Interpretation of NSA criteria for calling a stroke code likely varied within and between sites, and inter‐rater reliability of physical signs was not assessed, which is a limitation of the data. Observed rates of stroke for alerts that did not conform to the NSA criteria suggest that clinical judgment remains valuable. Final diagnoses were assigned by the stroke programs, and reliability of this assessment was not evaluated. Sites were not asked to use a specific categorization scheme to group final diagnoses. This analysis was limited to stroke centers with existing infrastructure to respond to stroke alerts and participated in an explicit quality‐improvement initiative on in‐hospital stroke response. Mimic and thrombolysis treatment rates may be different for hospitals without this stroke expertise.

Clinical uncertainty as to final diagnosis was addressed with the inclusion of confidence intervals accounting for potential misdiagnosis of the events in the categories of possible TIA or in the cases where the final diagnosis was unknown. Other studies have categorized TIA versus an alternative diagnosis as stroke mimic, and so our methodology is expected to yield a conservative estimate of the stroke mimic rate. Delirium is often a multifactorial phenomenon, so there may be an element of overlap between this category and other more specific mimic etiologies such as infection, hypotension, metabolic, or medication effect.

This initiative did not have the ability to assess the false negative rate of stroke team activation (failure to identify stroke symptoms in time for acute evaluation). It is not possible to calculate the sensitivity of stroke alerts in each center or conclude the optimal rate of false alarms. The finding of inter‐institutional variability in stroke alerts due to true brain ischemia could be explained by differences in staff education, systematic differences in the patient populations cared for among hospitals, or variation in institutional acceptance of having activated the stroke response team for cases with lower pretest probability of stroke. Sensitivity of alert criteria is more important than specificity, given the consequences of missing a potentially treatable emergent condition.

In conclusion, in this multi‐institution analysis of in‐hospital stroke alerts, a substantial proportion of in‐hospital strokes received thrombolytic therapy. Almost half of stroke alerts will not be for stroke or TIA. For many patients in our study, a change in neurologic status represented a harbinger of a change in general medical condition (hemorrhage, hypotension, hypoglycemia, or respiratory failure). Rapid response systems used for stroke in the hospital need to be trained and prepared to respond to a variety of acute medical conditions that extend beyond ischemic stroke.