- Email: [email protected]
Postpartum venous thromboembolism readmissions in the United States
SMFM Papers
ajog.org
Postpartum venous thromboembolism readmissions in the United States Timothy Wen, MD, MPH; Jason D. Wright, MD; Dena Goffman, MD; Mary E. D’Alton, MD; William J. Mack, MD, MS; Frank J. Attenello, MD, MS; Alexander M. Friedman, MD, MPH
BACKGROUND: There are limited data on when postpartum read-
missions for thromboembolism occur after delivery hospitalizations on a population basis in the United States. OBJECTIVE: We sought to characterize risk factors for and timing of postpartum venous thromboembolism readmission after delivery hospitalization discharge. STUDY DESIGN: The Healthcare Cost and Utilization Project Nationwide Readmissions Database for calendar years 2013 and 2014 was used to perform a retrospective cohort study evaluating risk for readmission for venous thromboembolism within 60 days of discharge from a delivery hospitalization. Risks for deep vein thrombosis and pulmonary embolism were individually assessed. Obstetric, medical, demographic, and hospital factors associated with postpartum readmission for venous thromboembolism were analyzed. Risk was characterized as odds ratios with 95% confidence intervals. Both unadjusted and adjusted analyses were performed. Adjusted analyses included relevant obstetric, medical, demographic, and hospital factors within logistic regression models. RESULTS: From Jan. 1 through Oct. 31 in 2013 and 2014, 6,269,641 delivery hospitalizations were included in the analysis. In all, 2975 cases of readmission for any venous thromboembolism were identified (4.7 per 10,000 delivery hospitalizations) including 1170 cases of deep vein thrombosis and 1805 cases of pulmonary embolism. In all, 69.6% of readmissions for any venous thromboembolism occurred within the first 20 days of discharge vs 22.3% and 8.0% at 21e40 and 41e60 days after discharge. Median times to readmission were 12.7, 14.0, and 11.7 days for venous thromboembolism, deep vein thrombosis, and pulmonary embolism, respectively. Women readmitted for any venous
Introduction Obstetric venous thromboembolism (VTE) is a leading cause of maternal mortality and severe morbidity in the United States.1 Pregnancy involves all 3 factors of Virchow triad2 and risk for events can be particularly high soon after delivery.3,4 In the United States, VTE during postpartum hospitalizations
Cite this article as: Wen T, Wright JD, Goffman D, et al. Postpartum venous thromboembolism readmissions in the United States. Am J Obstet Gynecol 2018;volume:x.ex-x.ex. 0002-9378/$36.00 ª 2018 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.ajog.2018.07.001
thromboembolism were more likely to have a history of venous thromboembolism (4.2% vs 0.3%, P < .01), to have had a cesarean delivery (54.4% vs 32.4%, P < .01), to have a thrombophilia (1.8% vs 0.4%, P < .01), to have had a longer delivery hospitalization of >3 days for vaginal delivery and >4 days for cesarean (18.0% vs 6.6%, P < .01), to have been diagnosed with gestational hypertension or preeclampsia (19.7% vs 8.2%, P < .01), and to have had postpartum hemorrhage with transfusion (2.6% vs 0.5%, P < .01). These factors retained significance in adjusted models. History of venous thromboembolism and hemorrhage with transfusion were associated with the largest odds of readmission (odds ratio, 9.5; 95% confidence interval, 6.6e13.6, and odds ratio, 3.6; 95% confidence interval, 2.4e5.5, respectively). Other factors associated with increased odds included thrombophilia (odds ratio, 2.0; 95% confidence interval, 1.2e3.5), cesarean delivery (odds ratio, 2.0; 95% confidence interval, 1.8e2.3), longer delivery hospitalization (odds ratio, 1.8; 95% confidence interval, 1.5e2.2), and preeclampsia or gestational hypertension (odds ratio, 2.0; 95% confidence interval, 1.6e2.4). CONCLUSION: While the majority of events occurred within 20 days of discharge, risk factors other than thrombophilia and prior venous thromboembolism were generally associated with modestly increased odds of events, and only a small proportion of readmissions occurred among women with thrombophilia and prior events. Our data demonstrate both the challenging nature and urgent need for further research to determine which clinical practices and interventions may reduce risk for venous thromboembolism readmissions on a population basis. Key words: obstetric thromboembolism, postpartum thromboembo-
lism, severe maternal morbidity
appears to be rising5 with the Centers for Disease Control and Prevention estimating the rate of thrombotic embolism during postpartum hospitalizations increased 169% from 1998 through 2009.6 Currently in the United States, obstetric patients at highest risk for VTE such as those with thrombophilia and prior VTE events receive routine pharmacologic postpartum thromboprophylaxis with unfractionated or low-molecular-weight heparin during delivery hospitalizations.7,8 Major guideline recommendations generally agree on continuation of pharmacologic prophylaxis after delivery hospitalization discharge for women with the highest
risk.9e11 However, little is known regarding timing and incidence of postpartum VTE readmission after delivery hospitalization discharge on a population basis in the United States. Knowledge of when events occur could be useful in patient counseling and management; if events are most proximal to discharge, recommendations to patients with risk factors could be made to ensure frequent ambulation and adequate hydration. Given that population-based postpartum readmission risk for thromboembolism is not well characterized, the purpose of this study was to evaluate risk for deep vein thrombosis (DVT) and pulmonary embolism (PE)
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e1
SMFM Papers
ajog.org
AJOG at a Glance Why was this study conducted? Given that population-based postpartum readmission risk for thromboembolism is not well characterized, the purpose of this study was to evaluate risk for deep vein thrombosis and pulmonary embolism after discharge home from a delivery hospitalization. Key findings In all, 69.6% of readmissions for any venous thromboembolism occurred within the first 20 days of discharge. Most risk factors were associated with only modestly increased risk for events. What does this add to what is known? This study demonstrates both the challenging nature and urgent need for further research to determine which clinical practices and interventions may reduce risk for venous thromboembolism readmissions on a population basis.
after discharge home from a delivery hospitalization.
Materials and Methods Data source The analysis was performed using the Healthcare Cost and Utilization Project Nationwide Readmissions Database (NRD) from 2013 and 2014. The NRD is an all-payer database with data
collected on a state level capable of tracking patients across hospital admissions within a state. The data can be used to create national estimates of readmissions for the insured and uninsured. The NRD includes public and community hospitals and academic medical centers.12 It has been used across a wide number of medical and surgical subspecialties to evaluate
FIGURE
VTE readmissions by days from delivery hospitalization discharge 50% 40% 30% 20% 10% 0% VTE
DVT
PE
Days aer discharge from delivery hospitalizaon: 1 to 10 11 to 20 21 to 30 31 to 40 41 to 50 51 to 60 Proportion of admissions for VTE, DVT, and PE occurring within 60 days of discharge from delivery hospitalization by 10-day periods (1e10, 11e20, 21e30, 31e40, 41e50, or 51e60 days). Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
1.e2 American Journal of Obstetrics & Gynecology MONTH 2018
readmission hospitalizations.13e15 Data in the NRD are weighted to provide national estimates with an estimated 35 million US discharges. In 2014, 22 geographically dispersed states contributed data to the NRD, accounting for 51% of US residents and 49% of all US hospitalizations.12 The Columbia University and University of Southern California institutional review boards granted exemptions given that the NRD is deidentified and publicly available.
Study population Index delivery hospitalizations were captured using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnoses codes 650 and V27.x. These criteria ascertain >95% of delivery hospitalizations.16 Women aged 15e54 years were included. VTE was defined as either acute DVT or PE. Diagnoses of acute PE were captured with ICD-9-CM codes 415.1x, 673.2x, 673.3x, and 673.8x.17 Diagnoses of acute DVT were captured with ICD-9-CM codes: 451.11, 451.19, 451.81, 451.83, 453.40, 453.41, 453.42, 453.8, 453.82, 453.83, 453.85, 453.87, 453.89, 453.9, 671.3x, 671.4x, and 671.5x.5,18,19 As the primary outcome of the study, we assessed risk for readmission with a new VTE diagnosis within 60 days of discharge from a delivery hospitalization. Women were determined to have a VTE readmission if they had a diagnosis of acute PE or DVT as the primary indication for postpartum readmission. Readmissions were identified using methodology provided by the Healthcare Cost and Utilization Project in the NRD. To account for multiple readmissions within 60 days, only the first readmission was included in the analysis. To avoid misclassification of historical vs acute diagnosis, women with a diagnosis of VTE during the delivery index hospitalization were excluded. Population weights from the NRD were applied to create national estimates. Because the NRD data sets are year-based and cannot be linked, only delivery hospitalizations where discharge occurred from Jan. 1 through Oct. 31 for each year were included; delivery hospitalizations during
SMFM Papers
ajog.org
TABLE 1
Demographic, medical, obstetric, and hospital factors associated with postpartum venous thromboembolism readmission No VTE readmission n
VTE readmission %
n
%
All patients Age, y 15e19
430,156
6.9%
156
5.3%
20e24
1,392,954
22.2%
635
21.3%
25e29
1,786,593
28.5%
798
26.8%
30e34
1,687,960
26.9%
768
25.8%
35e39
783,852
12.5%
427
14.4%
40e54
185,151
3.0%
191
6.4%
Lowest
1,647,152
26.3%
994
33.4%
Low
1,622,522
25.9%
822
27.6%
High
1,555,027
24.8%
709
23.8%
Highest
1,382,722
22.1%
424
14.3%
Missing
59,242
1.0%
26
0.9%
Medicare
48,656
0.7%
70
2.2%
Medicaid
3,179,159
42.5%
1645
50.7%
Private
3,894,056
52.0%
1392
42.9%
121,257
1.6%
35
1.1%
Median household ZIP code income quartile
Insurance status
Self-pay No charge Other Thrombophilia Pregestational diabetes
a
4660
0.1%
239,195
3.2%
101
3.1%
22,569
0.4%
54
1.8%
0.1%
64,059
1.0%
62
2.1%
342,263
5.5%
232
7.8%
18,920
0.3%
125
4.2%
Hypertensive diseases of pregnancy
511,957
8.2%
586
19.7%
Chronic hypertension
110,840
1.8%
116
3.9%
5,643,868
90.1%
2273
76.4%
2,032,431
32.4%
1618
54.4%
415,668
6.6%
535
18.0%
Postpartum hemorrhage with transfusion
29,137
0.5%
78
2.6%
Infectionc
23,060
0.4%
39
1.3%
119,626
1.9%
113
3.8%
Metro nonteaching
1,980,584
31.6%
852
28.6%
Metro teaching
3,632,218
58.0%
1784
60.0%
653,863
10.4%
339
11.4%
Tobacco History of VTE Hypertension
None Cesarean delivery Longer length of stay
b
Multiple gestation Hospital teaching
Nonmetro
Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
(continued)
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e3
SMFM Papers
ajog.org
TABLE 1
Demographic, medical, obstetric, and hospital factors associated with postpartum venous thromboembolism readmission (continued) No VTE readmission
VTE readmission
n
%
n
%
1,829,849
29.2%
724
24.4%
Large fringe metro counties
1,597,650
25.5%
701
23.7%
Medium metro counties
1,317,821
21.1%
679
22.9%
Small metro counties
608,707
9.7%
347
11.7%
Micropolitan (<50,000 population)
537,838
8.6%
321
10.8%
Not metro/micropolitan
366,704
5.9%
194
6.5%
835,760
13.3%
415
14.0%
Medium
1,751,182
27.9%
804
27.0%
Large
3,679,724
58.7%
1755
59.0%
Hospital location Large central metro counties
Hospital bed size Small
VTE, venous thromboembolism. a Data not presented given cell size <10; b >4 d for Cesarean delivery and >3 d for vaginal delivery; c Endometritis, pyelonephritis, pneumonia, sepsis, systemic inflammatory response syndrome. Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
November and December were not included because readmissions for the subsequent 60 days could not be fully ascertained. To characterize risk for women who typically do not receive postpartum VTE prophylaxis with postpartum heparin on discharge we repeated the analysis restricting the cohort to lower-risk women without a history of VTE events and thrombophilia.
Patient and hospital characteristics Demographic, medical, obstetric, and hospital factors in the NRD potentially associated with VTE risk from the index delivery hospitalization were analyzed. Demographic factors included payer (Medicaid, private, Medicare, self-pay, other), maternal age (15e19, 20e24, 25e29, 30e34, 35e39, and 40e54 years), and median household income quartile based on ZIP code. Medical factors measured during the index hospitalization included thrombophilia, pregestational diabetes, tobacco use, personal history of VTE, and chronic hypertension. Obstetric factors during the index admission included mode of delivery (cesarean vs vaginal),
hypertensive diseases of pregnancy (preeclampsia or gestational hypertension), postpartum hemorrhage with transfusion, infection (endometritis, pyelonephritis, pneumonia, sepsis, or systemic inflammatory response syndrome), multiple gestation, and delivery hospitalization stay. Hospital stays for delivery hospitalizations were dichotomized into longer and shorter stays; longer hospital stay was defined as >4 days for cesarean delivery and >3 days for vaginal delivery. Hospital factors included hospital bed size, teaching vs nonteaching status, and location based on the National Center for Health Statistics Urban-Rural Classification Scheme for Counties.12 Bed size classification was defined using region of the United States, the urban-rural designation of the hospital, and teaching status.12
Statistical analysis Risk for readmission with acute VTE was characterized in 10-day intervals for up to 60 days after delivery hospitalization discharge. Readmissions individually for DVT and PE diagnoses were then individually characterized in the same way up to 60 days after discharge. Univariable
1.e4 American Journal of Obstetrics & Gynecology MONTH 2018
analyses between risk for VTE readmission and hospital, demographic, and obstetric/medical risk factors were performed with Fisher exact test or c2 test as appropriate. Multivariable surveyadjusted logistic regression analyses were conducted to assess the relationship between patient, hospital, and demographic factors and the primary outcome of 60-day readmission for VTE. Odds ratios (OR) and 95% confidence intervals (CI) were used to estimate the effect of factors on odds for VTE readmission. Analysis was conducted using software (SAS 9.4; SAS Institute Inc, Cary, NC) with significance set at a P < .05.
Sensitivity analysis Because VTE is relatively rare in obstetric populations, small rates of misclassification may lead to relatively low positive predictive values for ICD-9-CM codes for VTE.20 We performed a sensitivity analysis restricted to the ICD9-CM codes with the highest positive predictive value (>75%) for an acute VTE event confirmed by chart review: 415.11, 453.8, 453.82, 453.83, 453.85, 453.87, 453.89, 451.11, 671.33, 673.21, 673.22, 673.24, and 673.83.20
SMFM Papers
ajog.org
TABLE 2
Adjusted and unadjusted odds of factors associated with postpartum venous thromboembolism readmission Unadjusted analysis OR
Adjusted analysis 95% CI
aOR
95% CI
0.80
0.59e1.09
0.74
0.54e1.01
Age, y 15e19 20e24
1.02
0.86e1.22
0.94
0.79e1.13
25e29
1.00
(Reference)
1.00
(Reference)
30e34
1.02
0.87e1.19
1.08
0.92e1.28
35e39
1.23
1.02e1.48
1.26
1.04e1.51
40e54
2.33
1.70e3.19
2.09
1.52e2.86
Lowest
1.97
1.60e2.42
1.74
1.40e2.15
Low
1.64
1.33e2.02
1.50
1.21e1.86
Median household ZIP code income quartile
High
1.49
1.21e1.83
1.43
1.17e1.76
Highest
1.00
(Reference)
1.00
(Reference)
4.14
2.72e6.32
2.71
1.78e4.14
Insurance status Medicare Medicaid
1.45
1.26e1.67
1.43
1.24e1.66
Private
1.00
(Reference)
1.00
(Reference)
Self-pay
0.87
0.50e1.51
0.91
0.52e1.59
No charge
2.75
0.41e18.33
2.76
0.41e18.63
1.18
0.81e1.71
1.18
0.82e1.71
Thrombophilia
Other
5.18
3.26e8.23
2.04
1.18e3.52
Pregestational diabetes
2.08
1.36e3.17
0.96
0.62e1.49
Tobacco
1.45
1.13e1.87
1.15
0.89e1.47
14.63
10.9e19.64
9.47
6.61e13.56
Hypertensive diseases of pregnancy
2.85
2.37e3.43
1.97
1.63e2.39
Chronic hypertension
2.63
1.98e3.50
1.69
1.26e2.27
2.47
2.17e2.81
2.04
1.79e2.33
History of VTE Hypertension
Cesarean delivery Longer length of stay
a
3.06
2.58e3.64
1.84
1.52e2.22
5.83
3.91e8.70
3.59
2.35e5.46
b
3.63
2.11e6.25
1.76
1.02e3.05
Multiple gestation
2.04
1.50e2.78
1.17
0.85e1.61
Postpartum hemorrhage with transfusion Infection
Hospital teaching Metro nonteaching
1.14
0.98e1.31
1.07
0.92e1.24
Metro teaching
1.19
0.96e1.48
0.93
0.70e1.24
Nonmetro
1.00
(Reference)
1.00
(Reference)
Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
Results Of 6,269,641 weighted deliveries from Jan. 1 through Oct. 31 in 2013 and 2014,
(continued)
2975 women were readmitted with a primary diagnosis of any VTE within 60 days of discharge (4.7 readmissions for
VTE per 10,000 delivery hospitalizations). Of these VTE readmissions 1170 were for DVT (1.9 per 10,000) and 1805
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e5
SMFM Papers
ajog.org
TABLE 2
Adjusted and unadjusted odds of factors associated with postpartum venous thromboembolism readmission (continued) Unadjusted analysis OR
Adjusted analysis 95% CI
aOR
95% CI
Hospital location Large central metro counties
1.00
(Reference)
1.00
(Reference)
Large fringe metro counties
1.12
0.93e1.36
1.28
1.06e1.54
Medium metro counties
1.32
1.1e1.58
1.43
1.19e1.72
Small metro counties
1.46
1.18e1.82
1.57
1.25e1.97
Micropolitan (<50,000 population)
1.53
1.22e1.92
1.69
1.26e2.27
Not metro/micropolitan
1.36
1.04e1.77
1.40
1.02e1.94
Small
1.00
(Reference)
1.00
(Reference)
Medium
0.95
0.77e1.17
1.07
0.92e1.24
Large
0.99
0.82e1.18
0.93
0.70e1.24
Hospital bed size
Adjusted model included all covariates listed in table. aOR, adjusted odds ratio; CI, confidence interval; OR, odds ratio; VTE, venous thromboembolism. >4 d for Cesarean delivery and >3 d for vaginal delivery; b Endometritis, pyelonephritis, pneumonia, sepsis, systemic inflammatory response syndrome. Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
a
were for PE (2.9 per 10,000). Readmissions were more likely proximal to discharge. For any VTE, 69.6% of readmissions occurred within the first 20 days of discharge, 22.3% occurred from 21e40 days after discharge, and 8.0% occurred 41e60 days after discharge. For PE, 70.2%, 21.7%, and 8.1% of readmissions occurred at 1e20, 21e40, and 41e60 days postdischarge. For DVT 68.8%, 23.2%, and 7.9% of readmissions occurred at 1e10, 11e30, and 31e60 days postdischarge (Figure). Median times to readmission were 12.7 days for any VTE (95% CI, 11.8e13.7), 14.0 days for DVT (95% CI, 12.7e15.3), and 11.7 days for PE (95% CI, 10.5e12.9). Readmitted patients had a mean length of stay of 4.0 days (95% CI, 3.8e4.2 days). Women who were readmitted for any VTE were more likely to have a history of VTE (4.2% vs 0.3%, P <.01), to have had a cesarean delivery (54.4% vs 32.4%, P <.01), to have had a prolonged delivery hospitalization (18.0% vs 6.6%, P < .01), to have a history of thrombophilia (1.8% vs 0.4%, P <.01), to have preeclampsia or gestational hypertension (19.7% vs 8.2%, P < .01), and to have had postpartum hemorrhage with transfusion (2.6% vs
0.5%, P < .01) than women not readmitted for VTE (Table 1). Demographically, women readmitted for VTE were more likely to be of advanced maternal age (20.8% vs 15.5%, P < .01), have Medicaid insurance (50.7% vs 42.5%, P < .01), and be from the ZIP code quartile with the lowest incomes (33.4% vs 26.3%, P < .01). In multivariable analysis (all variables were included in the adjusted model), many of the risk factors for any VTE readmission retained significance (Table 2). History of VTE and hemorrhage with transfusion were associated with the largest odds of readmission (OR, 9.5; 95% CI, 6.6e13.6, and OR, 3.6; 95% CI, 2.4e5.5, respectively). Other factors associated with increased odds included thrombophilia (OR, 2.0; 95% CI, 1.2e3.5), cesarean delivery (OR, 2.0; 95% CI, 1.8e2.3), longer length of stay during index hospitalization (OR, 1.8; 95% CI, 1.5e2.2), and hypertensive diseases of pregnancy (OR, 2.0; 95% CI, 1.6e2.4). Demographic factors associated with increased risk included maternal age 40e54 years with age 25e29 years as a reference (OR, 2.1; 95% CI, 1.5e2.9), being in the bottom ZIP code income quartile compared to
1.e6 American Journal of Obstetrics & Gynecology MONTH 2018
the top (OR, 1.7; 95% CI, 1.4e2.2), and Medicaid insurance compared to private (OR, 1.4; 95% CI, 1.2e1.7). Hospitals located in micropolitan and small metro counties had highest odds of VTE readmission (OR, 1.7; 95% CI, 1.3e2.3, and OR, 1.6; 95% CI, 1.3e2.0, respectively, with large central metro counties as a reference). When the adjusted analysis was restricted to women without prior VTE or a history of thrombophilia many risk factors were associated with increased odds of similar magnitude as the primary analysis (Table 3). Factors with the largest increased risk included maternal age 40e54 years (OR, 2.18; 95% CI, 1.58e3.01), hypertensive diseases of pregnancy (OR, 2.01; 95% CI, 1.65e2.45), cesarean delivery (OR, 2.05; 95% CI, 1.80e2.35), and postpartum hemorrhage with transfusion (OR, 3.46; 95% CI, 2.27e5.28). In the sensitivity analysis restricted to ICD-9-CM codes with highest sensitivity for VTE, 1478 events were diagnosed. Similar to the primary analysis, demographic factors associated significantly with increased risk for VTE readmission included advanced maternal age, ZIP code quartile with the lowest median income, and Medicaid
SMFM Papers
ajog.org
TABLE 3
Adjusted and unadjusted odds for postpartum venous thromboembolism for patients without thrombophilia or prior venous thromboembolism Unadjusted analysis
Adjusted analysis
OR
95% CI
aOR
95% CI
0.81
0.59e1.11
0.71
0.51e0.98
Age, y 15e19 20e24
1.02
0.86e1.22
0.94
0.79e1.13
25e29
1.00
(Reference)
1.00
(Reference)
30e34
1.00
0.84e1.18
1.09
0.92e1.29
35e39
1.22
1.01e1.47
1.28
1.06e1.55
40e54
2.33
1.69e3.21
2.18
1.58e3.01
1.98
1.61e2.42
1.77
1.43e2.19
Median household ZIP code income quartile Lowest Low
1.66
1.35e2.04
1.51
1.21e1.86
High
1.51
1.24e1.86
1.46
1.19e1.79
Highest
1.00
(Reference)
1.00
(Reference)
Insurance status Medicare
4.05
2.61e6.27
3.01
1.95e4.66
Medicaid
1.42
1.23e1.64
1.41
1.21e1.64
Private
1.00
(Reference)
1.00
(Reference)
Self-pay
0.90
0.51e1.57
0.94
0.54e1.64
No charge
2.84
0.42e18.93
2.83
0.42e19.12
Other
1.19
0.83e1.71
1.20
0.84e1.73
Pregestational diabetes
1.88
1.18e2.99
0.89
0.55e1.43
Tobacco
1.37
1.05e1.78
1.06
0.81e1.39
Hypertensive diseases of pregnancy
2.90
2.40e3.51
2.01
1.65e2.45
Chronic hypertension
2.57
1.91e3.47
1.72
1.25e2.35
2.44
2.14e2.79
2.05
1.80e2.35
3.13
2.63e3.72
1.96
1.62e2.36
5.64
3.77e8.45
3.46
2.27e5.28
Hypertension
Cesarean delivery Longer length of stay
a
Postpartum hemorrhage with transfusion b
3.57
2.03e6.28
1.77
0.99e3.14
Multiple gestation
2.05
1.50e2.82
1.19
0.85e1.65
Metro nonteaching
1.12
0.97e1.30
1.07
0.92e1.24
Metro teaching
1.21
0.97e1.50
0.94
0.70e1.26
Nonmetro
1.00
(Reference)
1.00
(Reference)
Infection
Hospital teaching
Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
insurance (Supplemental Table 1). Medical factors included history of VTE, thrombophilia, hypertensive diseases of pregnancy, cesarean delivery, longer length of stay during the delivery
(continued)
hospitalization, and postpartum hemorrhage with transfusion. In the adjusted analysis, history of VTE was associated with the highest odds of readmission (OR, 7.83; 95% CI, 4.5e13.8)
(Supplemental Table 2). Other factors associated with significantly increased odds in the adjusted sensitivity analysis included maternal age 40e54 years (OR, 2.7; 95% CI, 1.7e4.3), lowest compared
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e7
SMFM Papers
ajog.org
TABLE 3
Adjusted and unadjusted odds for postpartum venous thromboembolism for patients without thrombophilia or prior venous thromboembolism (continued) Unadjusted analysis
Adjusted analysis
OR
95% CI
aOR
95% CI
Large central metro counties
1.00
(Reference)
1.00
(Reference)
Hospital location Large fringe metro counties
1.12
0.93e1.36
1.28
1.06e1.55
Medium metro counties
1.33
1.11e1.60
1.46
1.21e1.77
Small metro counties
1.51
1.21e1.87
1.62
1.28e2.04
Micropolitan (<50,000 population)
1.54
1.22e1.94
1.75
1.29e2.37
Not metro/micropolitan
1.34
1.02e1.76
1.37
0.97e1.94
Small
1.00
(Reference)
1.00
(Reference)
Medium
0.95
0.76e1.18
0.92
0.74e1.15
Large
0.97
0.81e1.17
0.92
0.74e1.15
Hospital bed size
Adjusted model included all covariates listed in table. aOR, adjusted odds ratio; CI, confidence interval; OR, odds ratio. >4 d for Cesarean delivery and >3 d for vaginal delivery; b Endometritis, pyelonephritis, pneumonia, sepsis, systemic inflammatory response syndrome. Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
a
to highest median income ZIP code quartile (OR, 1.5; 95% CI, 1.1e2.0), Medicaid compared to private insurance (OR, 1.41; 95% CI, 1.1e1.8), cesarean delivery (OR, 2.4; 95% CI, 1.9e2.8), longer length of stay (OR, 1.9; 95% CI, 1.5e2.5), postpartum hemorrhage with transfusion (OR, 2.3; 95% CI, 1.2e4.5), and infection (OR, 2.5; 95% CI, 1.2e5.1%).
Comment Obstetric VTE is a leading cause of maternal mortality and severe morbidity.1,19,21,22 Findings from this analysis address the knowledge gap of when postpartum VTE readmissions occur after delivery hospitalizations on a population basis in the United States. This analysis demonstrated that risk is highest in the first 10 days after discharge and that most events occur within the first 3 weeks. These findings align with international studies demonstrating highest risk soon after delivery.23e27 Clinical implications of this study include that it may be reasonable to counsel women to be vigilant and report VTE symptoms on discharge and to continue regular ambulation and
adequate hydration when arriving home from the hospital. In evaluating risk factors for postpartum VTE readmission, the presence of prior events was associated with the largest increase in risk. This risk supports current recommendations across major society guidelines supporting postdischarge pharmacologic prophylaxis with low-molecular-weight or unfractionated heparin.9e11,28 However, readmissions occurring among women with prior events and thrombophilia— the highest risk group—accounted for <6% of cases. Other risk factors such as advanced maternal age, cesarean delivery, longer length of stay, hemorrhage with transfusion, and infection are all relatively common conditions that were associated with increased odds of VTE readmission in our model. While recommendations from the United Kingdom support longer courses of postpartum pharmacologic prophylaxis with low-molecular-weight heparin for women with these common risk factors,10 clear evidence supporting which patients may be most likely to benefit from extended prophylaxis is lacking. That many risk factors were broadly
1.e8 American Journal of Obstetrics & Gynecology MONTH 2018
distributed across the population and associated with only modestly increased risk underscores the difficulty of devising clinical strategies to reduce likelihood of these outcomes. That significant differences in risk were noted based on socioeconomic status, geographic location, and insurance status suggests the possibility that differences in care, rather than obstetric and medical risk factors alone, may play a role in outcomes. Strengths of this study include a large nationally representative database designed to analyze hospital readmissions, allowing an adjusted analysis for relatively rare VTE readmissions with multiple risk factors. The validity of our findings on the temporal distribution of readmissions was enhanced by the fact that both the primary and the sensitivity analyses demonstrated similar results. While misclassification and under-ascertainment are always concerns with administrative diagnosis codes, the sensitivity analysis for risk factors associated with VTE readmission restricted to higher sensitivity codes demonstrated similar results to the primary model in the univariate and adjusted analyses.
SMFM Papers
ajog.org Limitations of the study include shortcomings inherent to administrative data, which provide a broad overview of population-based risk, but lack many important clinical details. For example, given the poor capture of obesity, we were unable to adjust for this characteristic. Additionally, we are unable to account for many important aspects of clinical management during the delivery hospitalization such as what type of VTE prophylaxis was used and for what duration. Another limitation of this database is that we cannot account for outpatient management of a VTE event if a readmission did not occur, nor can we account for fatal PE events occurring outside the hospital. There may be provider variation in when to admit patients for DVT events, for example, and we are not able to capture these diagnoses. This database does not include data on outpatient pharmacy receipts or medications prescribed. We are not able to assess which patients may have received pharmacologic prophylaxis, at which dose, and for what duration. Other limitations of the database include that we could not track patients across calendar years and for that reason limited analyses to January through October in 2013 and 2014, and that the database is state-based, so if a patient delivered in a state different from where a readmission occurred we would not have linked the hospitalizations. Finally, we note that high-risk diagnoses may be underascertained by the diagnosis codes used, and a larger proportion of patients with VTE readmissions may be identified clinically. In interpreting these data, it is important to note that we presumed primary VTE diagnoses to be more likely to be valid and only these cases were included in our study; that secondary VTE diagnoses were not included may lead to underestimation of VTE readmissions and bias the study toward hospitalizations with VTE identified on an outpatient basis rather than during a readmission. In conclusions, this study provides novel epidemiologic data by characterizing time intervals from delivery discharge to readmission for a nationally representative population in the United
States. While the majority of events occurred within 20 days of discharge, risk factors other than thrombophilia and prior VTE were generally associated with modestly increased odds of events, and only a small proportion of readmissions occurred among women with thrombophilia and prior events. Our data demonstrate both the challenging nature and urgent need for further research to determine which clinical practices and interventions may reduce risk for VTE readmissions on a population basis. n References 1. Creanga AA, Syverson C, Seed K, Callaghan WM. Pregnancy-related mortality in the United States, 2011-2013. Obstet Gynecol 2017;130:366–73. 2. Bourjeily G, Paidas M, Khalil H, RoseneMontella K, Rodger M. Pulmonary embolism in pregnancy. Lancet 2010;375:500–12. 3. Abdul Sultan A, Grainge MJ, West J, Fleming KM, Nelson-Piercy C, Tata LJ. Impact of risk factors on the timing of first postpartum venous thromboembolism: a population-based cohort study from England. Blood 2014;124: 2872–80. 4. Heit JA, Kobbervig CE, James AH, Petterson TM, Bailey KR, Melton LJ III. Trends in the incidence of venous thromboembolism during pregnancy or postpartum: a 30-year population-based study. Ann Intern Med 2005;143:697–706. 5. Ghaji N, Boulet SL, Tepper N, Hooper WC. Trends in venous thromboembolism among pregnancy-related hospitalizations, United States, 1994-2009. Am J Obstet Gynecol 2013;209:433.e431–8. 6. Callaghan WM, Creanga AA, Kuklina EV. Severe maternal morbidity among delivery and postpartum hospitalizations in the United States. Obstet Gynecol 2012;120:1029–36. 7. Friedman AM, Ananth CV, Lu YS, D’Alton ME, Wright JD. Underuse of postcesarean thromboembolism prophylaxis. Obstet Gynecol 2013;122:1197–204. 8. Friedman AM, Ananth CV, Prendergast E, Chauhan SP, D’Alton ME, Wright JD. Thromboembolism incidence and prophylaxis during vaginal delivery hospitalizations. Am J Obstet Gynecol 2015;212:221.e1–12. 9. James A; Committee on Practice Bulletins— Obstetrics. Thromboembolism in pregnancy. Practice bulletin no. 123. Obstet Gynecol 2011;118:718–29. 10. Royal College of Obstetricians and Gynecologists. Thrombosis and embolism during pregnancy and the puerperium, reducing the risk. Green-Top guideline no. 37a. April 2015. Available at: https://www.rcog.org.uk/ globalassets/documents/guidelines/gtg-37a.pdf. Accessed July 25, 2018.
11. Bates SM, Greer IA, Middeldorp S, et al. VTE, thrombophilia, antithrombotic therapy, and pregnancy: antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest 2012;141(Suppl):e691736S. 12. Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project (HCUP). Introduction to the HCIP Nationwide Readmissions Database (NRD) 2010-2015. November 2017. Available at: https://www.hcup-us.ahrq. gov/db/nation/nrd/Introduction_NRD_2010-2015. pdf. Accessed March 1, 2018. 13. Kurtz SM, Lau EC, Ong KL, Adler EM, Kolisek FR, Manley MT. Which clinical and patient factors influence the national economic burden of hospital readmissions after total joint arthroplasty? Clin Orthop Relat Res 2017;475: 2926–37. 14. Vejpongsa P, Bhise V, Charitakis K, et al. Early readmissions after transcatheter and surgical aortic valve replacement. Catheter Cardiovasc Interv 2017;90:662–70. 15. Poojary P, Saha A, Chauhan K, et al. Predictors of hospital readmissions for ulcerative colitis in the United States: a national database study. Inflamm Bowel Dis 2017;23:347–56. 16. Kuklina E, Whiteman M, Hillis S, Jameieson D, Meikle S, Posner S. An enhanced method for identifying obstetric deliveries: implications for estimating maternal morbidity. Matern Child Health J 2008;12:469–77. 17. Centers for Disease Control and Prevention. Severe Maternal Morbidity. Available at: https://www.cdc.gov/reproductivehealth/ maternalinfanthealth/severematernalmorbidity. html Accessed March 1, 2018. 18. Tamariz L, Harkins T, Nair V. A systematic review of validated methods for identifying venous thromboembolism using administrative and claims data. Pharmacoepidemiol Drug Saf 2012;21(Suppl):154–62. 19. James AH, Jamison MG, Brancazio LR, Myers ER. Venous thromboembolism during pregnancy and the postpartum period: incidence, risk factors, and mortality. Am J Obstet Gynecol 2006;194:1311–5. 20. White RH, Brickner LA, Scannell KA. ICD-9CM codes poorly identified venous thromboembolism during pregnancy. J Clin Epidemiol 2004;57:985–8. 21. Clark SL, Belfort MA, Dildy GA, Herbst MA, Meyers JA, Hankins GD. Maternal death in the 21st century: causes, prevention, and relationship to cesarean delivery. Am J Obstet Gynecol 2008;199:36.e1–5;discussion 91-2. e7-11. 22. Clark SL, Christmas JT, Frye DR, Meyers JA, Perlin JB. Maternal mortality in the United States: predictability and the impact of protocols on fatal postcesarean pulmonary embolism and hypertension-related intracranial hemorrhage. Am J Obstet Gynecol 2014;211: 32.e31–9. 23. Morris J, Algert C, Roberts C. Incidence and risk factors for pulmonary embolism in the
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e9
SMFM Papers postpartum period. J Thromb Haemost 2010;8: 998–1003. 24. Heit JA, Kobbervig CE, James AH, Petterson TM, Bailey KR, Melton LJ. Trends in the incidence of venous thromboembolism during pregnancy or postpartum: a 30-year population-based study. Ann Intern Med 2005;143:697–706. 25. Virkus RA, Løkkegaard ECL, Bergholt T, Mogensen U, Langhoff-Roos J, Lidegaard Ø. Venous thromboembolism in pregnant and puerperal women in Denmark 1995e2005: a national cohort study. Obstet Gynecol Surv 2011;66:753–5. 26. Jacobsen AF, Skjeldestad FE, Sandset PM. Incidence and risk patterns of venous thromboembolism in pregnancy and
ajog.org puerperium—a register-based case-control study. Am J Obstet Gynecol 2008;198:233.e1–7. 27. Sultan AA, Grainge MJ, West J, Fleming KM, Nelson-Piercy C, Tata LJ. Impact of risk factors on the timing of first postpartum venous thromboembolism: a population-based cohort study from England. Blood 2014;124:2872–80. 28. D’Alton ME, Friedman AM, Smiley RM, et al. National Partnership for Maternal Safety: consensus bundle on venous thromboembolism. Obstet Gynecol 2016;128:688–98.
Author and article information From the Department of Obstetrics and Gynecology, College of Physicians and Surgeons, Columbia University,
1.e10 American Journal of Obstetrics & Gynecology MONTH 2018
New York, NY (Drs Wen, Wright, Goffman, D’Alton, and Friedman), and Department of Neurological Surgery, Keck School of Medicine, Los Angeles, CA (Drs Mack and Attenello). Received Feb. 11, 2018; revised June 26, 2018; accepted July 2, 2018. Dr Friedman is supported by a career development award (K08HD082287) from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health. Disclosure: Dr Wright has served as a consultant for Tesaro and Clovis Oncology. The remaining authors report no conflict of interest. Presented at the annual meeting of the Society for Maternal-Fetal Medicine, Dallas, TX, February 2, 2018. Corresponding author: Alexander M. Friedman, MD, MPH. [email protected]
SMFM Papers
ajog.org
SUPPLEMENTAL TABLE 1
Factors associated with postpartum venous thromboembolism readmission based on high sensitivity codes No VTE readmission n
VTE readmission %
n
%
All patients Age, y 15e19
430,263
6.9%
49
3.3%
20e24
1,393,317
22.2%
272
18.4%
25e29
1,787,015
28.5%
376
25.4%
30e34
1,688,318
26.9%
410
27.8%
35e39
784,027
12.5%
252
17.1%
40e54
185,223
3.0%
119
8.1%
Lowest
1,647,665
26.3%
480
32.5%
Low
1,622,956
25.9%
389
26.3%
High
1,555,396
24.8%
340
23.0%
Highest
1,382,887
22.1%
260
17.6%
Median household ZIP code income quartile
Insurance status Medicare
40,617
0.7%
32
2.2%
Medicaid
2,648,238
42.3%
692
47.1%
Private
3,261,815
52.2%
663
45.1%
99,935
1.6%
22
1.5%
Self-pay No charge
3950
0.1%
0
0.0%
199,736
3.2%
60
4.1%
Thrombophilia
22,605
0.4%
19
1.3%
Pregestational diabetes
64,099
1.0%
23
1.5%
342,371
5.5%
124
8.4%
18,993
0.3%
52
3.5%
Hypertensive diseases of pregnancy
512,203
8.2%
340
23.0%
Chronic hypertension
110,893
1.8%
63
4.3%
Other
Tobacco History of VTE Hypertension
None
5,645,066
90.1%
1074
72.7%
2,033,174
32.4%
875
59.2%
415,910
6.6%
293
19.8%
29,187
0.5%
28
1.9%
23,073
0.4%
26
1.8%
119,678
1.9%
61
4.1%
Metro nonteaching
1,981,035
31.6%
401
27.1%
Metro teaching
3,633,096
58.0%
906
61.3%
654,031
10.4%
171
11.6%
Cesarean delivery Longer length of staya Postpartum hemorrhage with transfusion Infection
b
Multiple gestation Hospital teaching
Nonmetro
Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
(continued)
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e11
SMFM Papers
ajog.org
SUPPLEMENTAL TABLE 1
Factors associated with postpartum venous thromboembolism readmission based on high sensitivity codes (continued) No VTE readmission
VTE readmission
n
%
n
%
Large central metro counties
1,830,253
29.2%
320
21.7%
Large fringe metro counties
1,597,980
25.5%
372
25.2%
Medium metro counties
Hospital location
1,318,156
21.1%
343
23.3%
Small metro counties
608,881
9.7%
173
11.7%
Micropolitan (<50,000 population)
537,998
8.6%
162
11.0%
Not metro/micropolitan
366,793
5.9%
105
7.1%
835,960
13.3%
215
14.5%
Medium
1,751,619
27.9%
367
24.9%
Large
3,680,584
58.7%
896
60.6%
Hospital bed size Small
VTE, venous thromboembolism. >4 d for Cesarean delivery and >3 d for vaginal delivery; b Endometritis, pyelonephritis, pneumonia, sepsis, systemic inflammatory response syndrome. Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
a
1.e12 American Journal of Obstetrics & Gynecology MONTH 2018
SMFM Papers
ajog.org
SUPPLEMENTAL TABLE 2
Adjusted and unadjusted odds for postpartum venous thromboembolism with high sensitivity codes Unadjusted analysis OR
Adjusted analysis 95% CI
aOR
95% CI
0.50
0.31e0.80
Age, y 15e19
0.54
0.34e0.86
20e24
0.92
0.70e1.20
0.85
0.65e1.11
25e29
1.00
(Reference)
1.00
(Reference)
30e34
1.13
0.89e1.45
1.21
0.94e1.54
35e39
1.53
1.15e2.02
1.55
1.16e2.06
40e54
3.06
1.94e4.84
2.70
1.72e4.25
Lowest
1.56
1.15e2.12
1.46
1.08e1.96
Low
1.25
0.92e1.71
1.19
0.88e1.61
Median household ZIP code income quartile
High
1.16
0.88e1.55
1.14
0.86e1.52
Highest
1.00
(Reference)
1.00
(Reference)
Medicare
3.86
2.06e7.22
2.52
1.34e4.73
Medicaid
1.29
1.04e1.59
1.41
1.14e1.75
Private
1.00
(Reference)
1.00
(Reference)
Self-pay
1.10
0.52e2.31
1.23
0.59e2.57
n/a
n/a
n/a
Insurance status
No charge
n/a 1.41
0.86e2.31
1.50
0.92e2.46
Thrombophilia
Other
3.53
1.52e8.22
1.40
0.52e3.72
Pregestational diabetes
1.52
0.83e2.78
0.60
0.32e1.13
Tobacco
1.60
1.16e2.22
1.30
0.93e1.81
12.17
7.65e19.37
7.83
4.45e13.76
Hypertensive diseases of pregnancy
3.49
2.68e4.54
2.36
1.80e3.10
Chronic hypertension
3.02
2.05e4.46
1.83
1.23e2.74
2.99
2.47e3.61
2.35
1.94e2.84
History of VTE Hypertension
Cesarean delivery Longer length of stay
a
3.46
2.75e4.37
1.93
1.51e2.47
4.20
2.25e7.85
2.34
1.22e4.50
b
4.89
2.36e10.14
2.45
1.17e5.10
Multiple gestation
2.22
1.48e3.35
1.13
0.73e1.75
Postpartum hemorrhage with transfusion Infection
Hospital teaching Metro nonteaching
1.22
0.98e1.52
1.15
0.92e1.43
Metro teaching
1.29
0.95e1.77
0.99
0.66e1.48
Nonmetro
1.00
(Reference)
1.00
(Reference)
Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
(continued)
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e13
SMFM Papers
ajog.org
SUPPLEMENTAL TABLE 2
Adjusted and unadjusted odds for postpartum venous thromboembolism with high sensitivity codes (continued) Unadjusted analysis OR
Adjusted analysis 95% CI
aOR
95% CI
Hospital location Large central metro counties
1.00
(Reference)
1.00
(Reference)
Large fringe metro counties
1.36
1.01e1.82
1.50
1.13e1.99
Medium metro counties
1.52
1.17e1.97
1.71
1.31e2.24
Small metro counties
1.66
1.22e2.27
1.93
1.40e2.66
Micropolitan (<50,000 population)
1.76
1.27e2.43
2.08
1.38e3.13
Not metro/micropolitan
1.67
1.17e2.38
1.85
1.18e2.92
Small
1.00
(Reference)
1.00
(Reference)
Medium
0.83
0.60e1.16
0.81
0.58e1.12
Large
0.97
0.75e1.27
0.88
0.68e1.15
Hospital bed size
Adjusted model included all covariates listed in table. aOR, adjusted odds ratio; CI, confidence interval; OR, odds ratio; VTE, venous thromboembolism. >4 d for cesarean delivery and >3 d for vaginal delivery; b Endometritis, pyelonephritis, pneumonia, sepsis, systemic inflammatory response syndrome. Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
a
1.e14 American Journal of Obstetrics & Gynecology MONTH 2018
ajog.org
Postpartum venous thromboembolism readmissions in the United States Timothy Wen, MD, MPH; Jason D. Wright, MD; Dena Goffman, MD; Mary E. D’Alton, MD; William J. Mack, MD, MS; Frank J. Attenello, MD, MS; Alexander M. Friedman, MD, MPH
BACKGROUND: There are limited data on when postpartum read-
missions for thromboembolism occur after delivery hospitalizations on a population basis in the United States. OBJECTIVE: We sought to characterize risk factors for and timing of postpartum venous thromboembolism readmission after delivery hospitalization discharge. STUDY DESIGN: The Healthcare Cost and Utilization Project Nationwide Readmissions Database for calendar years 2013 and 2014 was used to perform a retrospective cohort study evaluating risk for readmission for venous thromboembolism within 60 days of discharge from a delivery hospitalization. Risks for deep vein thrombosis and pulmonary embolism were individually assessed. Obstetric, medical, demographic, and hospital factors associated with postpartum readmission for venous thromboembolism were analyzed. Risk was characterized as odds ratios with 95% confidence intervals. Both unadjusted and adjusted analyses were performed. Adjusted analyses included relevant obstetric, medical, demographic, and hospital factors within logistic regression models. RESULTS: From Jan. 1 through Oct. 31 in 2013 and 2014, 6,269,641 delivery hospitalizations were included in the analysis. In all, 2975 cases of readmission for any venous thromboembolism were identified (4.7 per 10,000 delivery hospitalizations) including 1170 cases of deep vein thrombosis and 1805 cases of pulmonary embolism. In all, 69.6% of readmissions for any venous thromboembolism occurred within the first 20 days of discharge vs 22.3% and 8.0% at 21e40 and 41e60 days after discharge. Median times to readmission were 12.7, 14.0, and 11.7 days for venous thromboembolism, deep vein thrombosis, and pulmonary embolism, respectively. Women readmitted for any venous
Introduction Obstetric venous thromboembolism (VTE) is a leading cause of maternal mortality and severe morbidity in the United States.1 Pregnancy involves all 3 factors of Virchow triad2 and risk for events can be particularly high soon after delivery.3,4 In the United States, VTE during postpartum hospitalizations
Cite this article as: Wen T, Wright JD, Goffman D, et al. Postpartum venous thromboembolism readmissions in the United States. Am J Obstet Gynecol 2018;volume:x.ex-x.ex. 0002-9378/$36.00 ª 2018 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.ajog.2018.07.001
thromboembolism were more likely to have a history of venous thromboembolism (4.2% vs 0.3%, P < .01), to have had a cesarean delivery (54.4% vs 32.4%, P < .01), to have a thrombophilia (1.8% vs 0.4%, P < .01), to have had a longer delivery hospitalization of >3 days for vaginal delivery and >4 days for cesarean (18.0% vs 6.6%, P < .01), to have been diagnosed with gestational hypertension or preeclampsia (19.7% vs 8.2%, P < .01), and to have had postpartum hemorrhage with transfusion (2.6% vs 0.5%, P < .01). These factors retained significance in adjusted models. History of venous thromboembolism and hemorrhage with transfusion were associated with the largest odds of readmission (odds ratio, 9.5; 95% confidence interval, 6.6e13.6, and odds ratio, 3.6; 95% confidence interval, 2.4e5.5, respectively). Other factors associated with increased odds included thrombophilia (odds ratio, 2.0; 95% confidence interval, 1.2e3.5), cesarean delivery (odds ratio, 2.0; 95% confidence interval, 1.8e2.3), longer delivery hospitalization (odds ratio, 1.8; 95% confidence interval, 1.5e2.2), and preeclampsia or gestational hypertension (odds ratio, 2.0; 95% confidence interval, 1.6e2.4). CONCLUSION: While the majority of events occurred within 20 days of discharge, risk factors other than thrombophilia and prior venous thromboembolism were generally associated with modestly increased odds of events, and only a small proportion of readmissions occurred among women with thrombophilia and prior events. Our data demonstrate both the challenging nature and urgent need for further research to determine which clinical practices and interventions may reduce risk for venous thromboembolism readmissions on a population basis. Key words: obstetric thromboembolism, postpartum thromboembo-
lism, severe maternal morbidity
appears to be rising5 with the Centers for Disease Control and Prevention estimating the rate of thrombotic embolism during postpartum hospitalizations increased 169% from 1998 through 2009.6 Currently in the United States, obstetric patients at highest risk for VTE such as those with thrombophilia and prior VTE events receive routine pharmacologic postpartum thromboprophylaxis with unfractionated or low-molecular-weight heparin during delivery hospitalizations.7,8 Major guideline recommendations generally agree on continuation of pharmacologic prophylaxis after delivery hospitalization discharge for women with the highest
risk.9e11 However, little is known regarding timing and incidence of postpartum VTE readmission after delivery hospitalization discharge on a population basis in the United States. Knowledge of when events occur could be useful in patient counseling and management; if events are most proximal to discharge, recommendations to patients with risk factors could be made to ensure frequent ambulation and adequate hydration. Given that population-based postpartum readmission risk for thromboembolism is not well characterized, the purpose of this study was to evaluate risk for deep vein thrombosis (DVT) and pulmonary embolism (PE)
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e1
SMFM Papers
ajog.org
AJOG at a Glance Why was this study conducted? Given that population-based postpartum readmission risk for thromboembolism is not well characterized, the purpose of this study was to evaluate risk for deep vein thrombosis and pulmonary embolism after discharge home from a delivery hospitalization. Key findings In all, 69.6% of readmissions for any venous thromboembolism occurred within the first 20 days of discharge. Most risk factors were associated with only modestly increased risk for events. What does this add to what is known? This study demonstrates both the challenging nature and urgent need for further research to determine which clinical practices and interventions may reduce risk for venous thromboembolism readmissions on a population basis.
after discharge home from a delivery hospitalization.
Materials and Methods Data source The analysis was performed using the Healthcare Cost and Utilization Project Nationwide Readmissions Database (NRD) from 2013 and 2014. The NRD is an all-payer database with data
collected on a state level capable of tracking patients across hospital admissions within a state. The data can be used to create national estimates of readmissions for the insured and uninsured. The NRD includes public and community hospitals and academic medical centers.12 It has been used across a wide number of medical and surgical subspecialties to evaluate
FIGURE
VTE readmissions by days from delivery hospitalization discharge 50% 40% 30% 20% 10% 0% VTE
DVT
PE
Days aer discharge from delivery hospitalizaon: 1 to 10 11 to 20 21 to 30 31 to 40 41 to 50 51 to 60 Proportion of admissions for VTE, DVT, and PE occurring within 60 days of discharge from delivery hospitalization by 10-day periods (1e10, 11e20, 21e30, 31e40, 41e50, or 51e60 days). Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
1.e2 American Journal of Obstetrics & Gynecology MONTH 2018
readmission hospitalizations.13e15 Data in the NRD are weighted to provide national estimates with an estimated 35 million US discharges. In 2014, 22 geographically dispersed states contributed data to the NRD, accounting for 51% of US residents and 49% of all US hospitalizations.12 The Columbia University and University of Southern California institutional review boards granted exemptions given that the NRD is deidentified and publicly available.
Study population Index delivery hospitalizations were captured using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnoses codes 650 and V27.x. These criteria ascertain >95% of delivery hospitalizations.16 Women aged 15e54 years were included. VTE was defined as either acute DVT or PE. Diagnoses of acute PE were captured with ICD-9-CM codes 415.1x, 673.2x, 673.3x, and 673.8x.17 Diagnoses of acute DVT were captured with ICD-9-CM codes: 451.11, 451.19, 451.81, 451.83, 453.40, 453.41, 453.42, 453.8, 453.82, 453.83, 453.85, 453.87, 453.89, 453.9, 671.3x, 671.4x, and 671.5x.5,18,19 As the primary outcome of the study, we assessed risk for readmission with a new VTE diagnosis within 60 days of discharge from a delivery hospitalization. Women were determined to have a VTE readmission if they had a diagnosis of acute PE or DVT as the primary indication for postpartum readmission. Readmissions were identified using methodology provided by the Healthcare Cost and Utilization Project in the NRD. To account for multiple readmissions within 60 days, only the first readmission was included in the analysis. To avoid misclassification of historical vs acute diagnosis, women with a diagnosis of VTE during the delivery index hospitalization were excluded. Population weights from the NRD were applied to create national estimates. Because the NRD data sets are year-based and cannot be linked, only delivery hospitalizations where discharge occurred from Jan. 1 through Oct. 31 for each year were included; delivery hospitalizations during
SMFM Papers
ajog.org
TABLE 1
Demographic, medical, obstetric, and hospital factors associated with postpartum venous thromboembolism readmission No VTE readmission n
VTE readmission %
n
%
All patients Age, y 15e19
430,156
6.9%
156
5.3%
20e24
1,392,954
22.2%
635
21.3%
25e29
1,786,593
28.5%
798
26.8%
30e34
1,687,960
26.9%
768
25.8%
35e39
783,852
12.5%
427
14.4%
40e54
185,151
3.0%
191
6.4%
Lowest
1,647,152
26.3%
994
33.4%
Low
1,622,522
25.9%
822
27.6%
High
1,555,027
24.8%
709
23.8%
Highest
1,382,722
22.1%
424
14.3%
Missing
59,242
1.0%
26
0.9%
Medicare
48,656
0.7%
70
2.2%
Medicaid
3,179,159
42.5%
1645
50.7%
Private
3,894,056
52.0%
1392
42.9%
121,257
1.6%
35
1.1%
Median household ZIP code income quartile
Insurance status
Self-pay No charge Other Thrombophilia Pregestational diabetes
a
4660
0.1%
239,195
3.2%
101
3.1%
22,569
0.4%
54
1.8%
0.1%
64,059
1.0%
62
2.1%
342,263
5.5%
232
7.8%
18,920
0.3%
125
4.2%
Hypertensive diseases of pregnancy
511,957
8.2%
586
19.7%
Chronic hypertension
110,840
1.8%
116
3.9%
5,643,868
90.1%
2273
76.4%
2,032,431
32.4%
1618
54.4%
415,668
6.6%
535
18.0%
Postpartum hemorrhage with transfusion
29,137
0.5%
78
2.6%
Infectionc
23,060
0.4%
39
1.3%
119,626
1.9%
113
3.8%
Metro nonteaching
1,980,584
31.6%
852
28.6%
Metro teaching
3,632,218
58.0%
1784
60.0%
653,863
10.4%
339
11.4%
Tobacco History of VTE Hypertension
None Cesarean delivery Longer length of stay
b
Multiple gestation Hospital teaching
Nonmetro
Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
(continued)
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e3
SMFM Papers
ajog.org
TABLE 1
Demographic, medical, obstetric, and hospital factors associated with postpartum venous thromboembolism readmission (continued) No VTE readmission
VTE readmission
n
%
n
%
1,829,849
29.2%
724
24.4%
Large fringe metro counties
1,597,650
25.5%
701
23.7%
Medium metro counties
1,317,821
21.1%
679
22.9%
Small metro counties
608,707
9.7%
347
11.7%
Micropolitan (<50,000 population)
537,838
8.6%
321
10.8%
Not metro/micropolitan
366,704
5.9%
194
6.5%
835,760
13.3%
415
14.0%
Medium
1,751,182
27.9%
804
27.0%
Large
3,679,724
58.7%
1755
59.0%
Hospital location Large central metro counties
Hospital bed size Small
VTE, venous thromboembolism. a Data not presented given cell size <10; b >4 d for Cesarean delivery and >3 d for vaginal delivery; c Endometritis, pyelonephritis, pneumonia, sepsis, systemic inflammatory response syndrome. Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
November and December were not included because readmissions for the subsequent 60 days could not be fully ascertained. To characterize risk for women who typically do not receive postpartum VTE prophylaxis with postpartum heparin on discharge we repeated the analysis restricting the cohort to lower-risk women without a history of VTE events and thrombophilia.
Patient and hospital characteristics Demographic, medical, obstetric, and hospital factors in the NRD potentially associated with VTE risk from the index delivery hospitalization were analyzed. Demographic factors included payer (Medicaid, private, Medicare, self-pay, other), maternal age (15e19, 20e24, 25e29, 30e34, 35e39, and 40e54 years), and median household income quartile based on ZIP code. Medical factors measured during the index hospitalization included thrombophilia, pregestational diabetes, tobacco use, personal history of VTE, and chronic hypertension. Obstetric factors during the index admission included mode of delivery (cesarean vs vaginal),
hypertensive diseases of pregnancy (preeclampsia or gestational hypertension), postpartum hemorrhage with transfusion, infection (endometritis, pyelonephritis, pneumonia, sepsis, or systemic inflammatory response syndrome), multiple gestation, and delivery hospitalization stay. Hospital stays for delivery hospitalizations were dichotomized into longer and shorter stays; longer hospital stay was defined as >4 days for cesarean delivery and >3 days for vaginal delivery. Hospital factors included hospital bed size, teaching vs nonteaching status, and location based on the National Center for Health Statistics Urban-Rural Classification Scheme for Counties.12 Bed size classification was defined using region of the United States, the urban-rural designation of the hospital, and teaching status.12
Statistical analysis Risk for readmission with acute VTE was characterized in 10-day intervals for up to 60 days after delivery hospitalization discharge. Readmissions individually for DVT and PE diagnoses were then individually characterized in the same way up to 60 days after discharge. Univariable
1.e4 American Journal of Obstetrics & Gynecology MONTH 2018
analyses between risk for VTE readmission and hospital, demographic, and obstetric/medical risk factors were performed with Fisher exact test or c2 test as appropriate. Multivariable surveyadjusted logistic regression analyses were conducted to assess the relationship between patient, hospital, and demographic factors and the primary outcome of 60-day readmission for VTE. Odds ratios (OR) and 95% confidence intervals (CI) were used to estimate the effect of factors on odds for VTE readmission. Analysis was conducted using software (SAS 9.4; SAS Institute Inc, Cary, NC) with significance set at a P < .05.
Sensitivity analysis Because VTE is relatively rare in obstetric populations, small rates of misclassification may lead to relatively low positive predictive values for ICD-9-CM codes for VTE.20 We performed a sensitivity analysis restricted to the ICD9-CM codes with the highest positive predictive value (>75%) for an acute VTE event confirmed by chart review: 415.11, 453.8, 453.82, 453.83, 453.85, 453.87, 453.89, 451.11, 671.33, 673.21, 673.22, 673.24, and 673.83.20
SMFM Papers
ajog.org
TABLE 2
Adjusted and unadjusted odds of factors associated with postpartum venous thromboembolism readmission Unadjusted analysis OR
Adjusted analysis 95% CI
aOR
95% CI
0.80
0.59e1.09
0.74
0.54e1.01
Age, y 15e19 20e24
1.02
0.86e1.22
0.94
0.79e1.13
25e29
1.00
(Reference)
1.00
(Reference)
30e34
1.02
0.87e1.19
1.08
0.92e1.28
35e39
1.23
1.02e1.48
1.26
1.04e1.51
40e54
2.33
1.70e3.19
2.09
1.52e2.86
Lowest
1.97
1.60e2.42
1.74
1.40e2.15
Low
1.64
1.33e2.02
1.50
1.21e1.86
Median household ZIP code income quartile
High
1.49
1.21e1.83
1.43
1.17e1.76
Highest
1.00
(Reference)
1.00
(Reference)
4.14
2.72e6.32
2.71
1.78e4.14
Insurance status Medicare Medicaid
1.45
1.26e1.67
1.43
1.24e1.66
Private
1.00
(Reference)
1.00
(Reference)
Self-pay
0.87
0.50e1.51
0.91
0.52e1.59
No charge
2.75
0.41e18.33
2.76
0.41e18.63
1.18
0.81e1.71
1.18
0.82e1.71
Thrombophilia
Other
5.18
3.26e8.23
2.04
1.18e3.52
Pregestational diabetes
2.08
1.36e3.17
0.96
0.62e1.49
Tobacco
1.45
1.13e1.87
1.15
0.89e1.47
14.63
10.9e19.64
9.47
6.61e13.56
Hypertensive diseases of pregnancy
2.85
2.37e3.43
1.97
1.63e2.39
Chronic hypertension
2.63
1.98e3.50
1.69
1.26e2.27
2.47
2.17e2.81
2.04
1.79e2.33
History of VTE Hypertension
Cesarean delivery Longer length of stay
a
3.06
2.58e3.64
1.84
1.52e2.22
5.83
3.91e8.70
3.59
2.35e5.46
b
3.63
2.11e6.25
1.76
1.02e3.05
Multiple gestation
2.04
1.50e2.78
1.17
0.85e1.61
Postpartum hemorrhage with transfusion Infection
Hospital teaching Metro nonteaching
1.14
0.98e1.31
1.07
0.92e1.24
Metro teaching
1.19
0.96e1.48
0.93
0.70e1.24
Nonmetro
1.00
(Reference)
1.00
(Reference)
Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
Results Of 6,269,641 weighted deliveries from Jan. 1 through Oct. 31 in 2013 and 2014,
(continued)
2975 women were readmitted with a primary diagnosis of any VTE within 60 days of discharge (4.7 readmissions for
VTE per 10,000 delivery hospitalizations). Of these VTE readmissions 1170 were for DVT (1.9 per 10,000) and 1805
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e5
SMFM Papers
ajog.org
TABLE 2
Adjusted and unadjusted odds of factors associated with postpartum venous thromboembolism readmission (continued) Unadjusted analysis OR
Adjusted analysis 95% CI
aOR
95% CI
Hospital location Large central metro counties
1.00
(Reference)
1.00
(Reference)
Large fringe metro counties
1.12
0.93e1.36
1.28
1.06e1.54
Medium metro counties
1.32
1.1e1.58
1.43
1.19e1.72
Small metro counties
1.46
1.18e1.82
1.57
1.25e1.97
Micropolitan (<50,000 population)
1.53
1.22e1.92
1.69
1.26e2.27
Not metro/micropolitan
1.36
1.04e1.77
1.40
1.02e1.94
Small
1.00
(Reference)
1.00
(Reference)
Medium
0.95
0.77e1.17
1.07
0.92e1.24
Large
0.99
0.82e1.18
0.93
0.70e1.24
Hospital bed size
Adjusted model included all covariates listed in table. aOR, adjusted odds ratio; CI, confidence interval; OR, odds ratio; VTE, venous thromboembolism. >4 d for Cesarean delivery and >3 d for vaginal delivery; b Endometritis, pyelonephritis, pneumonia, sepsis, systemic inflammatory response syndrome. Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
a
were for PE (2.9 per 10,000). Readmissions were more likely proximal to discharge. For any VTE, 69.6% of readmissions occurred within the first 20 days of discharge, 22.3% occurred from 21e40 days after discharge, and 8.0% occurred 41e60 days after discharge. For PE, 70.2%, 21.7%, and 8.1% of readmissions occurred at 1e20, 21e40, and 41e60 days postdischarge. For DVT 68.8%, 23.2%, and 7.9% of readmissions occurred at 1e10, 11e30, and 31e60 days postdischarge (Figure). Median times to readmission were 12.7 days for any VTE (95% CI, 11.8e13.7), 14.0 days for DVT (95% CI, 12.7e15.3), and 11.7 days for PE (95% CI, 10.5e12.9). Readmitted patients had a mean length of stay of 4.0 days (95% CI, 3.8e4.2 days). Women who were readmitted for any VTE were more likely to have a history of VTE (4.2% vs 0.3%, P <.01), to have had a cesarean delivery (54.4% vs 32.4%, P <.01), to have had a prolonged delivery hospitalization (18.0% vs 6.6%, P < .01), to have a history of thrombophilia (1.8% vs 0.4%, P <.01), to have preeclampsia or gestational hypertension (19.7% vs 8.2%, P < .01), and to have had postpartum hemorrhage with transfusion (2.6% vs
0.5%, P < .01) than women not readmitted for VTE (Table 1). Demographically, women readmitted for VTE were more likely to be of advanced maternal age (20.8% vs 15.5%, P < .01), have Medicaid insurance (50.7% vs 42.5%, P < .01), and be from the ZIP code quartile with the lowest incomes (33.4% vs 26.3%, P < .01). In multivariable analysis (all variables were included in the adjusted model), many of the risk factors for any VTE readmission retained significance (Table 2). History of VTE and hemorrhage with transfusion were associated with the largest odds of readmission (OR, 9.5; 95% CI, 6.6e13.6, and OR, 3.6; 95% CI, 2.4e5.5, respectively). Other factors associated with increased odds included thrombophilia (OR, 2.0; 95% CI, 1.2e3.5), cesarean delivery (OR, 2.0; 95% CI, 1.8e2.3), longer length of stay during index hospitalization (OR, 1.8; 95% CI, 1.5e2.2), and hypertensive diseases of pregnancy (OR, 2.0; 95% CI, 1.6e2.4). Demographic factors associated with increased risk included maternal age 40e54 years with age 25e29 years as a reference (OR, 2.1; 95% CI, 1.5e2.9), being in the bottom ZIP code income quartile compared to
1.e6 American Journal of Obstetrics & Gynecology MONTH 2018
the top (OR, 1.7; 95% CI, 1.4e2.2), and Medicaid insurance compared to private (OR, 1.4; 95% CI, 1.2e1.7). Hospitals located in micropolitan and small metro counties had highest odds of VTE readmission (OR, 1.7; 95% CI, 1.3e2.3, and OR, 1.6; 95% CI, 1.3e2.0, respectively, with large central metro counties as a reference). When the adjusted analysis was restricted to women without prior VTE or a history of thrombophilia many risk factors were associated with increased odds of similar magnitude as the primary analysis (Table 3). Factors with the largest increased risk included maternal age 40e54 years (OR, 2.18; 95% CI, 1.58e3.01), hypertensive diseases of pregnancy (OR, 2.01; 95% CI, 1.65e2.45), cesarean delivery (OR, 2.05; 95% CI, 1.80e2.35), and postpartum hemorrhage with transfusion (OR, 3.46; 95% CI, 2.27e5.28). In the sensitivity analysis restricted to ICD-9-CM codes with highest sensitivity for VTE, 1478 events were diagnosed. Similar to the primary analysis, demographic factors associated significantly with increased risk for VTE readmission included advanced maternal age, ZIP code quartile with the lowest median income, and Medicaid
SMFM Papers
ajog.org
TABLE 3
Adjusted and unadjusted odds for postpartum venous thromboembolism for patients without thrombophilia or prior venous thromboembolism Unadjusted analysis
Adjusted analysis
OR
95% CI
aOR
95% CI
0.81
0.59e1.11
0.71
0.51e0.98
Age, y 15e19 20e24
1.02
0.86e1.22
0.94
0.79e1.13
25e29
1.00
(Reference)
1.00
(Reference)
30e34
1.00
0.84e1.18
1.09
0.92e1.29
35e39
1.22
1.01e1.47
1.28
1.06e1.55
40e54
2.33
1.69e3.21
2.18
1.58e3.01
1.98
1.61e2.42
1.77
1.43e2.19
Median household ZIP code income quartile Lowest Low
1.66
1.35e2.04
1.51
1.21e1.86
High
1.51
1.24e1.86
1.46
1.19e1.79
Highest
1.00
(Reference)
1.00
(Reference)
Insurance status Medicare
4.05
2.61e6.27
3.01
1.95e4.66
Medicaid
1.42
1.23e1.64
1.41
1.21e1.64
Private
1.00
(Reference)
1.00
(Reference)
Self-pay
0.90
0.51e1.57
0.94
0.54e1.64
No charge
2.84
0.42e18.93
2.83
0.42e19.12
Other
1.19
0.83e1.71
1.20
0.84e1.73
Pregestational diabetes
1.88
1.18e2.99
0.89
0.55e1.43
Tobacco
1.37
1.05e1.78
1.06
0.81e1.39
Hypertensive diseases of pregnancy
2.90
2.40e3.51
2.01
1.65e2.45
Chronic hypertension
2.57
1.91e3.47
1.72
1.25e2.35
2.44
2.14e2.79
2.05
1.80e2.35
3.13
2.63e3.72
1.96
1.62e2.36
5.64
3.77e8.45
3.46
2.27e5.28
Hypertension
Cesarean delivery Longer length of stay
a
Postpartum hemorrhage with transfusion b
3.57
2.03e6.28
1.77
0.99e3.14
Multiple gestation
2.05
1.50e2.82
1.19
0.85e1.65
Metro nonteaching
1.12
0.97e1.30
1.07
0.92e1.24
Metro teaching
1.21
0.97e1.50
0.94
0.70e1.26
Nonmetro
1.00
(Reference)
1.00
(Reference)
Infection
Hospital teaching
Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
insurance (Supplemental Table 1). Medical factors included history of VTE, thrombophilia, hypertensive diseases of pregnancy, cesarean delivery, longer length of stay during the delivery
(continued)
hospitalization, and postpartum hemorrhage with transfusion. In the adjusted analysis, history of VTE was associated with the highest odds of readmission (OR, 7.83; 95% CI, 4.5e13.8)
(Supplemental Table 2). Other factors associated with significantly increased odds in the adjusted sensitivity analysis included maternal age 40e54 years (OR, 2.7; 95% CI, 1.7e4.3), lowest compared
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e7
SMFM Papers
ajog.org
TABLE 3
Adjusted and unadjusted odds for postpartum venous thromboembolism for patients without thrombophilia or prior venous thromboembolism (continued) Unadjusted analysis
Adjusted analysis
OR
95% CI
aOR
95% CI
Large central metro counties
1.00
(Reference)
1.00
(Reference)
Hospital location Large fringe metro counties
1.12
0.93e1.36
1.28
1.06e1.55
Medium metro counties
1.33
1.11e1.60
1.46
1.21e1.77
Small metro counties
1.51
1.21e1.87
1.62
1.28e2.04
Micropolitan (<50,000 population)
1.54
1.22e1.94
1.75
1.29e2.37
Not metro/micropolitan
1.34
1.02e1.76
1.37
0.97e1.94
Small
1.00
(Reference)
1.00
(Reference)
Medium
0.95
0.76e1.18
0.92
0.74e1.15
Large
0.97
0.81e1.17
0.92
0.74e1.15
Hospital bed size
Adjusted model included all covariates listed in table. aOR, adjusted odds ratio; CI, confidence interval; OR, odds ratio. >4 d for Cesarean delivery and >3 d for vaginal delivery; b Endometritis, pyelonephritis, pneumonia, sepsis, systemic inflammatory response syndrome. Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
a
to highest median income ZIP code quartile (OR, 1.5; 95% CI, 1.1e2.0), Medicaid compared to private insurance (OR, 1.41; 95% CI, 1.1e1.8), cesarean delivery (OR, 2.4; 95% CI, 1.9e2.8), longer length of stay (OR, 1.9; 95% CI, 1.5e2.5), postpartum hemorrhage with transfusion (OR, 2.3; 95% CI, 1.2e4.5), and infection (OR, 2.5; 95% CI, 1.2e5.1%).
Comment Obstetric VTE is a leading cause of maternal mortality and severe morbidity.1,19,21,22 Findings from this analysis address the knowledge gap of when postpartum VTE readmissions occur after delivery hospitalizations on a population basis in the United States. This analysis demonstrated that risk is highest in the first 10 days after discharge and that most events occur within the first 3 weeks. These findings align with international studies demonstrating highest risk soon after delivery.23e27 Clinical implications of this study include that it may be reasonable to counsel women to be vigilant and report VTE symptoms on discharge and to continue regular ambulation and
adequate hydration when arriving home from the hospital. In evaluating risk factors for postpartum VTE readmission, the presence of prior events was associated with the largest increase in risk. This risk supports current recommendations across major society guidelines supporting postdischarge pharmacologic prophylaxis with low-molecular-weight or unfractionated heparin.9e11,28 However, readmissions occurring among women with prior events and thrombophilia— the highest risk group—accounted for <6% of cases. Other risk factors such as advanced maternal age, cesarean delivery, longer length of stay, hemorrhage with transfusion, and infection are all relatively common conditions that were associated with increased odds of VTE readmission in our model. While recommendations from the United Kingdom support longer courses of postpartum pharmacologic prophylaxis with low-molecular-weight heparin for women with these common risk factors,10 clear evidence supporting which patients may be most likely to benefit from extended prophylaxis is lacking. That many risk factors were broadly
1.e8 American Journal of Obstetrics & Gynecology MONTH 2018
distributed across the population and associated with only modestly increased risk underscores the difficulty of devising clinical strategies to reduce likelihood of these outcomes. That significant differences in risk were noted based on socioeconomic status, geographic location, and insurance status suggests the possibility that differences in care, rather than obstetric and medical risk factors alone, may play a role in outcomes. Strengths of this study include a large nationally representative database designed to analyze hospital readmissions, allowing an adjusted analysis for relatively rare VTE readmissions with multiple risk factors. The validity of our findings on the temporal distribution of readmissions was enhanced by the fact that both the primary and the sensitivity analyses demonstrated similar results. While misclassification and under-ascertainment are always concerns with administrative diagnosis codes, the sensitivity analysis for risk factors associated with VTE readmission restricted to higher sensitivity codes demonstrated similar results to the primary model in the univariate and adjusted analyses.
SMFM Papers
ajog.org Limitations of the study include shortcomings inherent to administrative data, which provide a broad overview of population-based risk, but lack many important clinical details. For example, given the poor capture of obesity, we were unable to adjust for this characteristic. Additionally, we are unable to account for many important aspects of clinical management during the delivery hospitalization such as what type of VTE prophylaxis was used and for what duration. Another limitation of this database is that we cannot account for outpatient management of a VTE event if a readmission did not occur, nor can we account for fatal PE events occurring outside the hospital. There may be provider variation in when to admit patients for DVT events, for example, and we are not able to capture these diagnoses. This database does not include data on outpatient pharmacy receipts or medications prescribed. We are not able to assess which patients may have received pharmacologic prophylaxis, at which dose, and for what duration. Other limitations of the database include that we could not track patients across calendar years and for that reason limited analyses to January through October in 2013 and 2014, and that the database is state-based, so if a patient delivered in a state different from where a readmission occurred we would not have linked the hospitalizations. Finally, we note that high-risk diagnoses may be underascertained by the diagnosis codes used, and a larger proportion of patients with VTE readmissions may be identified clinically. In interpreting these data, it is important to note that we presumed primary VTE diagnoses to be more likely to be valid and only these cases were included in our study; that secondary VTE diagnoses were not included may lead to underestimation of VTE readmissions and bias the study toward hospitalizations with VTE identified on an outpatient basis rather than during a readmission. In conclusions, this study provides novel epidemiologic data by characterizing time intervals from delivery discharge to readmission for a nationally representative population in the United
States. While the majority of events occurred within 20 days of discharge, risk factors other than thrombophilia and prior VTE were generally associated with modestly increased odds of events, and only a small proportion of readmissions occurred among women with thrombophilia and prior events. Our data demonstrate both the challenging nature and urgent need for further research to determine which clinical practices and interventions may reduce risk for VTE readmissions on a population basis. n References 1. Creanga AA, Syverson C, Seed K, Callaghan WM. Pregnancy-related mortality in the United States, 2011-2013. Obstet Gynecol 2017;130:366–73. 2. Bourjeily G, Paidas M, Khalil H, RoseneMontella K, Rodger M. Pulmonary embolism in pregnancy. Lancet 2010;375:500–12. 3. Abdul Sultan A, Grainge MJ, West J, Fleming KM, Nelson-Piercy C, Tata LJ. Impact of risk factors on the timing of first postpartum venous thromboembolism: a population-based cohort study from England. Blood 2014;124: 2872–80. 4. Heit JA, Kobbervig CE, James AH, Petterson TM, Bailey KR, Melton LJ III. Trends in the incidence of venous thromboembolism during pregnancy or postpartum: a 30-year population-based study. Ann Intern Med 2005;143:697–706. 5. Ghaji N, Boulet SL, Tepper N, Hooper WC. Trends in venous thromboembolism among pregnancy-related hospitalizations, United States, 1994-2009. Am J Obstet Gynecol 2013;209:433.e431–8. 6. Callaghan WM, Creanga AA, Kuklina EV. Severe maternal morbidity among delivery and postpartum hospitalizations in the United States. Obstet Gynecol 2012;120:1029–36. 7. Friedman AM, Ananth CV, Lu YS, D’Alton ME, Wright JD. Underuse of postcesarean thromboembolism prophylaxis. Obstet Gynecol 2013;122:1197–204. 8. Friedman AM, Ananth CV, Prendergast E, Chauhan SP, D’Alton ME, Wright JD. Thromboembolism incidence and prophylaxis during vaginal delivery hospitalizations. Am J Obstet Gynecol 2015;212:221.e1–12. 9. James A; Committee on Practice Bulletins— Obstetrics. Thromboembolism in pregnancy. Practice bulletin no. 123. Obstet Gynecol 2011;118:718–29. 10. Royal College of Obstetricians and Gynecologists. Thrombosis and embolism during pregnancy and the puerperium, reducing the risk. Green-Top guideline no. 37a. April 2015. Available at: https://www.rcog.org.uk/ globalassets/documents/guidelines/gtg-37a.pdf. Accessed July 25, 2018.
11. Bates SM, Greer IA, Middeldorp S, et al. VTE, thrombophilia, antithrombotic therapy, and pregnancy: antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest 2012;141(Suppl):e691736S. 12. Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project (HCUP). Introduction to the HCIP Nationwide Readmissions Database (NRD) 2010-2015. November 2017. Available at: https://www.hcup-us.ahrq. gov/db/nation/nrd/Introduction_NRD_2010-2015. pdf. Accessed March 1, 2018. 13. Kurtz SM, Lau EC, Ong KL, Adler EM, Kolisek FR, Manley MT. Which clinical and patient factors influence the national economic burden of hospital readmissions after total joint arthroplasty? Clin Orthop Relat Res 2017;475: 2926–37. 14. Vejpongsa P, Bhise V, Charitakis K, et al. Early readmissions after transcatheter and surgical aortic valve replacement. Catheter Cardiovasc Interv 2017;90:662–70. 15. Poojary P, Saha A, Chauhan K, et al. Predictors of hospital readmissions for ulcerative colitis in the United States: a national database study. Inflamm Bowel Dis 2017;23:347–56. 16. Kuklina E, Whiteman M, Hillis S, Jameieson D, Meikle S, Posner S. An enhanced method for identifying obstetric deliveries: implications for estimating maternal morbidity. Matern Child Health J 2008;12:469–77. 17. Centers for Disease Control and Prevention. Severe Maternal Morbidity. Available at: https://www.cdc.gov/reproductivehealth/ maternalinfanthealth/severematernalmorbidity. html Accessed March 1, 2018. 18. Tamariz L, Harkins T, Nair V. A systematic review of validated methods for identifying venous thromboembolism using administrative and claims data. Pharmacoepidemiol Drug Saf 2012;21(Suppl):154–62. 19. James AH, Jamison MG, Brancazio LR, Myers ER. Venous thromboembolism during pregnancy and the postpartum period: incidence, risk factors, and mortality. Am J Obstet Gynecol 2006;194:1311–5. 20. White RH, Brickner LA, Scannell KA. ICD-9CM codes poorly identified venous thromboembolism during pregnancy. J Clin Epidemiol 2004;57:985–8. 21. Clark SL, Belfort MA, Dildy GA, Herbst MA, Meyers JA, Hankins GD. Maternal death in the 21st century: causes, prevention, and relationship to cesarean delivery. Am J Obstet Gynecol 2008;199:36.e1–5;discussion 91-2. e7-11. 22. Clark SL, Christmas JT, Frye DR, Meyers JA, Perlin JB. Maternal mortality in the United States: predictability and the impact of protocols on fatal postcesarean pulmonary embolism and hypertension-related intracranial hemorrhage. Am J Obstet Gynecol 2014;211: 32.e31–9. 23. Morris J, Algert C, Roberts C. Incidence and risk factors for pulmonary embolism in the
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e9
SMFM Papers postpartum period. J Thromb Haemost 2010;8: 998–1003. 24. Heit JA, Kobbervig CE, James AH, Petterson TM, Bailey KR, Melton LJ. Trends in the incidence of venous thromboembolism during pregnancy or postpartum: a 30-year population-based study. Ann Intern Med 2005;143:697–706. 25. Virkus RA, Løkkegaard ECL, Bergholt T, Mogensen U, Langhoff-Roos J, Lidegaard Ø. Venous thromboembolism in pregnant and puerperal women in Denmark 1995e2005: a national cohort study. Obstet Gynecol Surv 2011;66:753–5. 26. Jacobsen AF, Skjeldestad FE, Sandset PM. Incidence and risk patterns of venous thromboembolism in pregnancy and
ajog.org puerperium—a register-based case-control study. Am J Obstet Gynecol 2008;198:233.e1–7. 27. Sultan AA, Grainge MJ, West J, Fleming KM, Nelson-Piercy C, Tata LJ. Impact of risk factors on the timing of first postpartum venous thromboembolism: a population-based cohort study from England. Blood 2014;124:2872–80. 28. D’Alton ME, Friedman AM, Smiley RM, et al. National Partnership for Maternal Safety: consensus bundle on venous thromboembolism. Obstet Gynecol 2016;128:688–98.
Author and article information From the Department of Obstetrics and Gynecology, College of Physicians and Surgeons, Columbia University,
1.e10 American Journal of Obstetrics & Gynecology MONTH 2018
New York, NY (Drs Wen, Wright, Goffman, D’Alton, and Friedman), and Department of Neurological Surgery, Keck School of Medicine, Los Angeles, CA (Drs Mack and Attenello). Received Feb. 11, 2018; revised June 26, 2018; accepted July 2, 2018. Dr Friedman is supported by a career development award (K08HD082287) from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health. Disclosure: Dr Wright has served as a consultant for Tesaro and Clovis Oncology. The remaining authors report no conflict of interest. Presented at the annual meeting of the Society for Maternal-Fetal Medicine, Dallas, TX, February 2, 2018. Corresponding author: Alexander M. Friedman, MD, MPH. [email protected]
SMFM Papers
ajog.org
SUPPLEMENTAL TABLE 1
Factors associated with postpartum venous thromboembolism readmission based on high sensitivity codes No VTE readmission n
VTE readmission %
n
%
All patients Age, y 15e19
430,263
6.9%
49
3.3%
20e24
1,393,317
22.2%
272
18.4%
25e29
1,787,015
28.5%
376
25.4%
30e34
1,688,318
26.9%
410
27.8%
35e39
784,027
12.5%
252
17.1%
40e54
185,223
3.0%
119
8.1%
Lowest
1,647,665
26.3%
480
32.5%
Low
1,622,956
25.9%
389
26.3%
High
1,555,396
24.8%
340
23.0%
Highest
1,382,887
22.1%
260
17.6%
Median household ZIP code income quartile
Insurance status Medicare
40,617
0.7%
32
2.2%
Medicaid
2,648,238
42.3%
692
47.1%
Private
3,261,815
52.2%
663
45.1%
99,935
1.6%
22
1.5%
Self-pay No charge
3950
0.1%
0
0.0%
199,736
3.2%
60
4.1%
Thrombophilia
22,605
0.4%
19
1.3%
Pregestational diabetes
64,099
1.0%
23
1.5%
342,371
5.5%
124
8.4%
18,993
0.3%
52
3.5%
Hypertensive diseases of pregnancy
512,203
8.2%
340
23.0%
Chronic hypertension
110,893
1.8%
63
4.3%
Other
Tobacco History of VTE Hypertension
None
5,645,066
90.1%
1074
72.7%
2,033,174
32.4%
875
59.2%
415,910
6.6%
293
19.8%
29,187
0.5%
28
1.9%
23,073
0.4%
26
1.8%
119,678
1.9%
61
4.1%
Metro nonteaching
1,981,035
31.6%
401
27.1%
Metro teaching
3,633,096
58.0%
906
61.3%
654,031
10.4%
171
11.6%
Cesarean delivery Longer length of staya Postpartum hemorrhage with transfusion Infection
b
Multiple gestation Hospital teaching
Nonmetro
Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
(continued)
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e11
SMFM Papers
ajog.org
SUPPLEMENTAL TABLE 1
Factors associated with postpartum venous thromboembolism readmission based on high sensitivity codes (continued) No VTE readmission
VTE readmission
n
%
n
%
Large central metro counties
1,830,253
29.2%
320
21.7%
Large fringe metro counties
1,597,980
25.5%
372
25.2%
Medium metro counties
Hospital location
1,318,156
21.1%
343
23.3%
Small metro counties
608,881
9.7%
173
11.7%
Micropolitan (<50,000 population)
537,998
8.6%
162
11.0%
Not metro/micropolitan
366,793
5.9%
105
7.1%
835,960
13.3%
215
14.5%
Medium
1,751,619
27.9%
367
24.9%
Large
3,680,584
58.7%
896
60.6%
Hospital bed size Small
VTE, venous thromboembolism. >4 d for Cesarean delivery and >3 d for vaginal delivery; b Endometritis, pyelonephritis, pneumonia, sepsis, systemic inflammatory response syndrome. Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
a
1.e12 American Journal of Obstetrics & Gynecology MONTH 2018
SMFM Papers
ajog.org
SUPPLEMENTAL TABLE 2
Adjusted and unadjusted odds for postpartum venous thromboembolism with high sensitivity codes Unadjusted analysis OR
Adjusted analysis 95% CI
aOR
95% CI
0.50
0.31e0.80
Age, y 15e19
0.54
0.34e0.86
20e24
0.92
0.70e1.20
0.85
0.65e1.11
25e29
1.00
(Reference)
1.00
(Reference)
30e34
1.13
0.89e1.45
1.21
0.94e1.54
35e39
1.53
1.15e2.02
1.55
1.16e2.06
40e54
3.06
1.94e4.84
2.70
1.72e4.25
Lowest
1.56
1.15e2.12
1.46
1.08e1.96
Low
1.25
0.92e1.71
1.19
0.88e1.61
Median household ZIP code income quartile
High
1.16
0.88e1.55
1.14
0.86e1.52
Highest
1.00
(Reference)
1.00
(Reference)
Medicare
3.86
2.06e7.22
2.52
1.34e4.73
Medicaid
1.29
1.04e1.59
1.41
1.14e1.75
Private
1.00
(Reference)
1.00
(Reference)
Self-pay
1.10
0.52e2.31
1.23
0.59e2.57
n/a
n/a
n/a
Insurance status
No charge
n/a 1.41
0.86e2.31
1.50
0.92e2.46
Thrombophilia
Other
3.53
1.52e8.22
1.40
0.52e3.72
Pregestational diabetes
1.52
0.83e2.78
0.60
0.32e1.13
Tobacco
1.60
1.16e2.22
1.30
0.93e1.81
12.17
7.65e19.37
7.83
4.45e13.76
Hypertensive diseases of pregnancy
3.49
2.68e4.54
2.36
1.80e3.10
Chronic hypertension
3.02
2.05e4.46
1.83
1.23e2.74
2.99
2.47e3.61
2.35
1.94e2.84
History of VTE Hypertension
Cesarean delivery Longer length of stay
a
3.46
2.75e4.37
1.93
1.51e2.47
4.20
2.25e7.85
2.34
1.22e4.50
b
4.89
2.36e10.14
2.45
1.17e5.10
Multiple gestation
2.22
1.48e3.35
1.13
0.73e1.75
Postpartum hemorrhage with transfusion Infection
Hospital teaching Metro nonteaching
1.22
0.98e1.52
1.15
0.92e1.43
Metro teaching
1.29
0.95e1.77
0.99
0.66e1.48
Nonmetro
1.00
(Reference)
1.00
(Reference)
Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
(continued)
MONTH 2018 American Journal of Obstetrics & Gynecology
1.e13
SMFM Papers
ajog.org
SUPPLEMENTAL TABLE 2
Adjusted and unadjusted odds for postpartum venous thromboembolism with high sensitivity codes (continued) Unadjusted analysis OR
Adjusted analysis 95% CI
aOR
95% CI
Hospital location Large central metro counties
1.00
(Reference)
1.00
(Reference)
Large fringe metro counties
1.36
1.01e1.82
1.50
1.13e1.99
Medium metro counties
1.52
1.17e1.97
1.71
1.31e2.24
Small metro counties
1.66
1.22e2.27
1.93
1.40e2.66
Micropolitan (<50,000 population)
1.76
1.27e2.43
2.08
1.38e3.13
Not metro/micropolitan
1.67
1.17e2.38
1.85
1.18e2.92
Small
1.00
(Reference)
1.00
(Reference)
Medium
0.83
0.60e1.16
0.81
0.58e1.12
Large
0.97
0.75e1.27
0.88
0.68e1.15
Hospital bed size
Adjusted model included all covariates listed in table. aOR, adjusted odds ratio; CI, confidence interval; OR, odds ratio; VTE, venous thromboembolism. >4 d for cesarean delivery and >3 d for vaginal delivery; b Endometritis, pyelonephritis, pneumonia, sepsis, systemic inflammatory response syndrome. Wen et al. Postpartum venous thromboembolism readmissions. Am J Obstet Gynecol 2018.
a
1.e14 American Journal of Obstetrics & Gynecology MONTH 2018