Abstract
Background:
Patients with hypertensive disorders of pregnancy have a high rate of postpartum readmission.
Objectives:
We evaluated whether type of antihypertensive medication prescribed at discharge was associated with postpartum readmission after a hypertensive disorder of pregnancy.
Study Design:
A retrospective cohort of 57,254 pregnancies with hypertensive disorders of pregnancy that began between 2012–2018 in the electronic obstetric database of Kaiser Permanente Northern California. Postpartum readmissions occurred within 6 weeks of discharge from the delivery hospitalization. Cox regression models evaluated associations between type of antihypertensive medication prescription at discharge (none, labetalol only, nifedipine only, two or more antihypertensive medications) with postpartum readmission, adjusted for type of hypertensive disorder of pregnancy, final inpatient systolic and diastolic blood pressures, age, body mass index, mode of delivery, insurance status, race/ethnicity, delivery facility, comorbidity score, smoking, preterm delivery, parity, and neighborhood deprivation index.
Results:
Among eligible patients with a hypertensive disorder of pregnancy, 1696 (3.0%) were readmitted within 6 weeks. Approximately 86% of patients were discharged without antihypertensive medication; among those discharged with a prescription for antihypertensive medication, most were prescribed either labetalol only (54%) or nifedipine only (30%). Unadjusted readmission risk was highest among patients discharged with labetalol only (7.6%), lower for patients discharged with nifedipine only (3.6%) or two or more antihypertensive medications (3.2%), and lowest for those discharged without a prescription for antihypertensive medication (2.5%). In adjusted models, compared to discharge without antihypertensive medication, discharge on labetalol only was associated with 63% (hazard ratio: 1.63; 95% confidence interval: 1.41, 1.88) greater incidence of postpartum readmission, while discharge on nifedipine only or on two or more antihypertensive medications were associated with 26% and 47% lower incidence of postpartum readmission, respectively (hazard ratios: 0.74 (95% confidence interval: 0.59, 0.93); 0.53 (95% confidence interval: 0.38, 0.74), respectively). There was no strong evidence to suggest that the effect of discharge antihypertensive medication type on incidence of readmission varied by race and ethnicity (interaction p=0.88). Results indicating an elevated risk for labetalol were consistent in models excluding patients with pre-pregnancy hypertension.
Conclusions:
Discharge on nifedipine alone or multiple antihypertensive medications (versus no medication) were associated with lower incidence of readmission, while discharge on labetalol alone was associated with elevated readmission incidence. Large-scale, prospective research to compare effectiveness of commonly prescribed hypertension medications at discharge is warranted.
Keywords: hypertension, labetalol, nifedipine, preeclampsia
Tweetable statement:
After a hypertensive disorder of pregnancy, postpartum discharge on labetalol was associated with higher risk of readmission, while discharge on nifedipine or multiple meds was associated with lower risk.
Introduction
Postpartum readmissions represent acute postpartum health problems1 and have been increasing.2 Readmission can be financially costly3 and physically disruptive to postpartum families. Hypertensive disorders of pregnancy (HDP) have been identified as an important predictor of potentially preventable postpartum readmission.2,3 Additional clinical and demographic factors linked to readmission among patients with HDP include older age,4–6 higher parity,5,7 Black (versus White) race,5,8,9 earlier gestational age at delivery,5,10 and greater HDP severity.3,4 Elevated postpartum blood pressure before discharge, including blood pressures below the American College of Obstetrics and Gynecology (ACOG) 2013 clinical guideline for initiating antihypertensive treatment (150/100 mm Hg)11 have also been linked to risk of postpartum readmission,5,6 though a specific lower threshold has not been identified.8,12,13
The effect of antihypertensive medication at discharge to prevent postpartum readmission after HDP or peripartum hypertension is unclear, with recent studies reporting both positive and null results.5,10,14 Additionally, it is unknown which antihypertensive medication is most effective to prevent postpartum readmission. Two recent studies suggest that risk of postpartum readmission may be greater for patients discharged on labetalol than nifedipine, two commonly prescribed antihypertensive medications.15,16 However, one study was conducted using claims data and could not incorporate information on inpatient blood pressures, parity, and body mass index;15 and the other was small (n=217 readmissions) and focused on a patient population that was more than 80% White,16 limiting generalizability.5,8,9 If discharge medication effectiveness varies, choice of antihypertensive medication may represent an actionable clinical target to reduce postpartum readmissions after HDP.
In a large integrated health care delivery system serving a racially and ethnically diverse patient population, we examined associations between type of discharge antihypertensive medication (none, nifedipine only, labetalol only, two or more antihypertensive medications) and risk of readmission among patients with HDP, adjusting for demographic and clinical factors including type of HDP and last inpatient postpartum blood pressure. Because HDP and readmission risks are characterized by racial disparities,9,17 we also explored the extent to which associations of discharge medication with postpartum readmission varied by race and ethnicity.
Methods
Kaiser Permanente Northern California (KPNC) is a large, integrated healthcare-delivery system providing medical care for approximately one-third of Northern California residents. Membership within KPNC is representative of the demographic characteristics of the region.18 KPNC is vertically integrated with all care provided in a closed system and captured in the electronic health record (EHR). Access to the EHR in an integrated health system allows us to collect high-quality clinical data including physician diagnoses and medical treatments during pregnancy.
This retrospective analysis included all pregnancies with a hypertensive diagnosis that began between 2012 and 2018; were discharged home after delivery; and were prescribed either no antihypertensive medication or nifedipine only, labetalol only, or multiple antihypertensive medications at discharge (n=57,254 pregnancies). Data collected during pregnancy, the delivery admission, and the first 6 weeks post-discharge were extracted from the EHR. This study was approved by the KPNC Institutional Review Board. The requirement for informed consent was waived.
Exposure: Medication
Using discharge instructions to continue antihypertensive medication as well as new prescriptions filled within one day post-discharge, we recorded use of any antihypertensive medication and specific medications (labetalol only, nifedipine only, or multiple antihypertensive medications). The multiple medications group included patients prescribed any combination of nifedipine, labetalol, or other classes of antihypertensive medication (83% were prescribed both nifedipine and labetalol, 11% were prescribed labetalol and another medication, and 6% were prescribed nifedipine and another medication).
Outcome: Readmission
We defined postpartum readmission as readmission to a KPNC facility within 6 weeks of discharge from the delivery hospitalization.
Covariates
Covariates were chosen using a directed acyclic graph (Supplemental Figure S1). Each hypertensive disorder (gestational hypertension, pre-existing hypertension, non-severe preeclampsia, pre-existing hypertension with superimposed preeclampsia, or severe preeclampsia) was classified based on International Classification of Diseases 9th and 10th revisions (ICD-9, −10) codes, except for severe preeclampsia, which was defined as meeting criteria for preeclampsia and relevant biochemical abnormalities (e.g., elevated liver function tests) following the ACOG diagnostic definition.19 We collected the final inpatient systolic and diastolic blood pressure measurement in the EHR, which was taken within 24 hours of discharge for >99% of participants. Systolic blood pressure was divided into 10-mm Hg categories (i.e., ≤120, 121–130, 131–140, 141–150, >150 mm Hg) and diastolic blood pressure was dichotomized at 80 mm Hg (i.e., ≤80, >80 mm Hg).
Maternal age at delivery, mode of delivery, delivery facility, parity, and type of insurance coverage was obtained from the EHR. We obtained maternal smoking during pregnancy using data from a self-administered Early Start Program Prenatal Substance Use Screening Questionnaire, which is completed at entry into prenatal care.20 Pre-pregnancy body mass index (BMI) was calculated using pre-pregnancy body weight (kilograms) divided by height (meters) squared. Pre-pregnancy body weight was assessed using clinically measured weight within 12 months before pregnancy (available for 90% of participants); if no measured weight was found , we used a self-reported pre-pregnancy weight or a pregnancy weight measured before 10 weeks’ gestation. Pre-pregnancy height was assessed using recorded height within the 4 years before pregnancy. BMI categories were created based on the World Health Organization International Classification.21
Preterm birth was defined as a livebirth <37 weeks’ gestation; gestational age at delivery, based on the best estimated delivery date assigned by obstetrical staff,22 was also extracted from the EHR.23,24 We previously validated obstetric estimate-based gestational age at delivery against early ultrasonographic examinations (gold standard) among both preterm and term infants.24
Neighborhood deprivation index is a composite variable based on census tract-level poverty, occupational and employment factors, education, and housing,25 determined based on maternal ZIP code during pregnancy; we standardized the index so that 0 represents a mean neighborhood and higher values represent neighborhoods with greater deprivation. Comorbidity Point Score, version 2 (COPS2), defined at delivery admission, integrates all patient diagnoses in the preceding 12 months into a continuous composite score representing burden of comorbidities.26,27
Maternal race and ethnicity, which are self-reported, were obtained from the EHR. We consider self-reported race and ethnicity as a proxy for social and structural factors that vary by societally determined race and ethnicity and affect lived experience, rather than a biological variable.
Statistical analysis
We used Cox proportional hazards regression models with days since discharge as the time scale to evaluate associations of antihypertensive medication prescription at discharge with incidence of postpartum readmission within 6 weeks. Patient death during follow-up (n=7) was the only censoring event. We used the missing-indicator approach to account for covariate missingness (≤6% for all covariates). Statistical models were adjusted for HDP type, final inpatient diastolic blood pressure (i.e., ≤80, >80 mm Hg), final inpatient systolic blood pressure (≤120, 121–130, 131–140, 141–150, >150 mm Hg), patient age (<25, 25–29, 30–34, ≥35 years), pre-pregnancy BMI (<25.0, 25.0–29.9, ≥30.0 kg/m2), mode of delivery (Cesarean versus vaginal), patient insurance status (Commercial, Medicaid/Medicare, Other), patient race and ethnicity (Asian/Pacific Islander, Black, Hispanic, Multiracial, White, Other, Unknown), delivery facility (16 facilities), delivery year, COPS2, smoking status (current smoker, nonsmoker, missing), preterm delivery, parity (nulliparous, primiparous, multiparous), and neighborhood deprivation index (<−1, −1 to <0, 0 to <1, ≥1).25,28 We additionally tested effect modification by race and ethnicity of the association between discharge antihypertensive medication and readmission risk. To assess the extent to which the effect is driven by patients with pre-pregnancy hypertension, we repeated the analysis excluding patients with pre-pregnancy hypertension. To reduce unmeasured confounding by indication, we also repeated the analysis with labetalol monotherapy as the reference group. The analysis was conducted in R version 4.1.2.
Results
Among 57,254 eligible patients, 1696 (3.0%) were readmitted within 6 weeks after discharge. Readmission was more likely among patients aged >35 years (35% of readmitted vs 24% of non-readmitted patients), with pre-pregnancy BMI ≥ 30.0 kg/m2 (43% vs 39%), who self-identified as non-Hispanic Black (13% vs 9%) or Asian/Pacific Islander (23% vs 19%), and who had a Cesarean delivery (44% vs 32%) (Table 1).
Table 1.
Demographic and clinical characteristics of patients with hypertensive disorders of pregnancy by postpartum readmission within 6 weeks post-discharge
| Overall (n=57,254) | Not readmitted (n=55,558, 97.0%) | Readmitted (n=1,696, 3.0%) | |
|---|---|---|---|
| Patient age, years, n(%) | |||
| < 25 | 9522 (16.6) | 9307 (16.8) | 215 (12.7) |
| 25 – 29 | 14933 (26.1) | 14605 (26.3) | 328 (19.3) |
| 30 – 34 | 18996 (33.2) | 18439 (33.2) | 557 (32.8) |
| ≥ 35 | 13803 (24.1) | 13207 (23.8) | 596 (35.1) |
| Pre-pregnancy BMI, kg/m2, n(%) | |||
| < 25.0 | 16860 (29.4) | 16368 (29.5) | 492 (29.0) |
| 25.0 – 29.9 | 15100 (26.4) | 14677 (26.4) | 423 (24.9) |
| ≥ 30.0 | 22476 (39.3) | 21745 (39.1) | 731 (43.1) |
| Missing | 2818 (4.9) | 2768 (5.0) | 50 (2.9) |
| Patient race/ethnicity, n(%) | |||
| Asian/Pacific Islander | 10824 (18.9) | 10438 (18.8) | 386 (22.8) |
| Black | 5096 (8.9) | 4882 (8.8) | 214 (12.6) |
| Hispanic | 14361 (25.1) | 13944 (25.1) | 417 (24.6) |
| Multiracial | 1772 (3.1) | 1719 (3.1) | 53 (3.1) |
| White | 23682 (41.4) | 23098 (41.6) | 584 (34.4) |
| Other | 835 (1.5) | 809 (1.5) | 26 (1.5) |
| Unknown or missing | 684 (1.2) | 668 (1.2) | 16 (0.9) |
| Parity | |||
| Nulliparous | 29333 (51.2) | 28487 (51.3) | 846 (49.9) |
| Primiparous | 15170 (26.5) | 14720 (26.5) | 450 (26.5) |
| Multiparous | 10198 (17.8) | 9877 (17.8) | 321 (18.9) |
| Missing | 2553 (4.5) | 2474 (4.5) | 79 (4.7) |
| Neighborhood Deprivation Index, n (%)1 | |||
| < −1 | 8178 (14.3) | 7946 (14.3) | 232 (13.7) |
| −1 to < 0 | 26738 (46.7) | 25895 (46.6) | 843 (49.7) |
| 0 to < 1 | 15518 (27.1) | 15084 (27.2) | 434 (25.6) |
| ≥ 1 | 6711 (11.7) | 6528 (11.7) | 183 (10.8) |
| Missing | 109 (0.2) | 105 (0.2) | 4 (0.2) |
| Patient insurance status, n(%) | |||
| Commercial | 49573 (86.6) | 48119 (86.6) | 1454 (85.7) |
| Medicare/Medicaid | 6328 (11.1) | 6125 (11.0) | 203 (12.0) |
| Other | 1353 (2.4) | 1314 (2.4) | 39 (2.3) |
| Current smoking, n (%) | |||
| Yes | 1671 (2.9) | 1616 (2.9) | 55 (3.2) |
| No | 52026 (90.9) | 50502 (90.9) | 1524 (89.9) |
| Missing | 3557 (6.2) | 3440 (6.2) | 117 (6.9) |
| Cesarean delivery, n (%) | 18767 (32.8) | 18015 (32.4) | 752 (44.3) |
| COPS2 score at admission, mean (SD) | 10.10 (3.39) | 10.09 (3.31) | 10.35 (5.32) |
| Preterm delivery, n (%) | 7308 (12.8) | 6989 (12.6) | 319 (18.8) |
| Hypertensive diagnosis | |||
| Gestational hypertension | 28410 (49.6) | 27885 (50.2) | 525 (31.0) |
| Pre-existing hypertension | 7582 (13.2) | 7331 (13.2) | 251 (14.8) |
| Non-severe pre-eclampsia | 9155 (16.0) | 8834 (15.9) | 321 (18.9) |
| Pre-existing hypertension with preeclampsia | 1359 (2.4) | 1290 (2.3) | 69 (4.1) |
| Severe pre-eclampsia | 10748 (18.8) | 10218 (18.4) | 530 (31.2) |
| Last inpatient systolic blood pressure, mm Hg | |||
| ≤ 120 | 16610 (29.0) | 16332 (29.4) | 278 (16.4) |
| 121–130 | 17383 (30.4) | 16988 (30.6) | 395 (23.3) |
| 131–140 | 15221 (26.6) | 14692 (26.4) | 529 (31.2) |
| 141–150 | 6529 (11.4) | 6146 (11.1) | 383 (22.6) |
| > 150 | 1511 (2.6) | 1400 (2.5) | 111 (6.5) |
| Last inpatient diastolic blood pressure, mm Hg | |||
| >80 | 20021 (35.0) | 19176 (34.5) | 845 (49.8) |
Higher values indicate greater neighborhood deprivation
Among patients with HDP, 86% were discharged without antihypertensive medication. Compared to patients not prescribed antihypertensive medication at discharge, those prescribed antihypertensive medication were more likely to have severe preeclampsia (55% vs 13%) and last inpatient systolic blood pressure >140 mm Hg (42% vs 10%). Patients prescribed versus not prescribed antihypertensive medication were more likely to be Black (13% vs 8%) or Asian/Pacific Islander (24% vs 18%), and less likely to be White (33% vs 43%); and were more likely to be ≥35 years old (35% vs 22%), to have BMI ≥30 kg/m2 (47% vs 38%), to have had a Cesarean delivery (51% vs 30%), and to have delivered preterm (38% vs 9%) (Table 2).
Table 2.
Characteristics of patients with hypertensive disorders of pregnancy by discharge antihypertensive medication
| Overall (n=57,254) | None (n=49,295) | Labetalol only (n=4,288) | Nifedipine only (n=2,426) | Two or more (n=1,245) | |
|---|---|---|---|---|---|
| Hypertensive diagnosis | |||||
| Gestational hypertension | 28410 (49.6) | 27941 (56.7) | 255 (5.9) | 192 (7.9) | 22 (1.8) |
| Pre-existing hypertension | 7582 (13.2) | 5836 (11.8) | 1205 (28.1) | 381 (15.7) | 160 (12.9) |
| Non-severe pre-eclampsia | 9155 (16.0) | 8351 (16.9) | 419 (9.8) | 323 (13.3) | 62 (5.0) |
| Pre-existing hypertension with preeclampsia | 1359 (2.4) | 786 (1.6) | 302 (7.0) | 155 (6.4) | 116 (9.3) |
| Severe pre-eclampsia | 10748 (18.8) | 6381 (12.9) | 2107 (49.1) | 1375 (56.7) | 885 (71.1) |
| Last systolic blood pressure, mm Hg, n (%) | |||||
| ≤120 | 16610 (29.0) | 16059 (32.6) | 300 (7.0) | 170 (7.0) | 81 (6.5) |
| 121–130 | 17383 (30.4) | 16019 (32.5) | 687 (16.0) | 475 (19.6) | 202 (16.2) |
| 131–140 | 15221 (26.6) | 12513 (25.4) | 1426 (33.3) | 882 (36.4) | 400 (32.1) |
| 141–150 | 6529 (11.4) | 4110 (8.3) | 1333 (31.1) | 711 (29.3) | 375 (30.1) |
| >150 | 1511 (2.6) | 594 (1.2) | 542 (12.6) | 188 (7.7) | 187 (15.0) |
| Last diastolic blood pressure >80 mm Hg, n (%) | 20021 (35.0) | 15034 (30.5) | 2594 (60.5) | 1594 (65.7) | 799 (64.2) |
| Patient age, years, n (%) | |||||
| < 25 | 9522 (16.6) | 8640 (17.5) | 464 (10.8) | 298 (12.3) | 120 (9.6) |
| 25 – 29 | 14933 (26.1) | 13322 (27.0) | 865 (20.2) | 518 (21.4) | 228 (18.3) |
| 30 – 34 | 18996 (33.2) | 16307 (33.1) | 1440 (33.6) | 832 (34.3) | 417 (33.5) |
| ≥ 35 | 13803 (24.1) | 11026 (22.4) | 1519 (35.4) | 778 (32.1) | 480 (38.6) |
| Pre-pregnancy BMI, kg/m2, n (%) | |||||
| < 25.0 | 16860 (29.4) | 15000 (30.4) | 948 (22.1) | 659 (27.2) | 253 (20.3) |
| 25.0 – 29.9 | 15100 (26.4) | 13097 (26.6) | 1072 (25.0) | 616 (25.4) | 315 (25.3) |
| ≥ 30.0 | 22476 (39.3) | 18769 (38.1) | 2047 (47.7) | 1035 (42.7) | 625 (50.2) |
| Missing | 2818 (4.9) | 2429 (4.9) | 221 (5.2) | 116 (4.8) | 52 (4.2) |
| Patient race/ethnicity, n (%) | |||||
| Asian/Pacific Islander | 10824 (18.9) | 8949 (18.2) | 1032 (24.1) | 554 (22.8) | 289 (23.2) |
| Black | 5096 (8.9) | 4026 (8.2) | 524 (12.2) | 295 (12.2) | 251 (20.2) |
| Hispanic | 14361 (25.1) | 12454 (25.3) | 1029 (24.0) | 592 (24.4) | 286 (23.0) |
| Multiracial | 1772 (3.1) | 1510 (3.1) | 134 (3.1) | 90 (3.7) | 38 (3.1) |
| White | 23682 (41.4) | 21092 (42.8) | 1428 (33.3) | 825 (34.0) | 337 (27.1) |
| Other | 835 (1.5) | 678 (1.4) | 91 (2.1) | 42 (1.7) | 24 (1.9) |
| Unknown or missing | 684 (1.2) | 586 (1.2) | 50 (1.2) | 28 (1.2) | 20 (1.6) |
| Parity | |||||
| Nulliparous | 29333 (51.2) | 25543 (51.8) | 1976 (46.1) | 1236 (50.9) | 578 (46.4) |
| Primiparous | 15170 (26.5) | 13127 (26.6) | 1146 (26.7) | 600 (24.7) | 297 (23.9) |
| Multiparous | 10198 (17.8) | 8459 (17.2) | 969 (22.6) | 464 (19.1) | 306 (24.6) |
| Missing | 2553 (4.5) | 2166 (4.4) | 197 (4.6) | 126 (5.2) | 64 (5.1) |
| Neighborhood Deprivation Index, n (%)1 | |||||
| < −1 | 8178 (14.3) | 7157 (14.5) | 510 (11.9) | 361 (14.9) | 150 (12.0) |
| −1 to < 0 | 26738 (46.7) | 23173 (47.0) | 1920 (44.8) | 1093 (45.1) | 552 (44.3) |
| 0 to < 1 | 15518 (27.1) | 13231 (26.8) | 1275 (29.7) | 650 (26.8) | 362 (29.1) |
| ≥ 1 | 6711 (11.7) | 5640 (11.4) | 572 (13.3) | 320 (13.2) | 179 (14.4) |
| Missing | 109 (0.2) | 94 (0.2) | 11 (0.3) | 2 (0.1) | 2 (0.2) |
| Patient insurance status, n (%) | |||||
| Commercial | 49573 (86.6) | 42741 (86.7) | 3684 (85.9) | 2099 (86.5) | 1049 (84.3) |
| Medicare/Medicaid | 6328 (11.1) | 5401 (11.0) | 492 (11.5) | 272 (11.2) | 163 (13.1) |
| Other | 1353 (2.4) | 1153 (2.3) | 112 (2.6) | 55 (2.3) | 33 (2.7) |
| Current smoking, n (%) | |||||
| Yes | 1671 (2.9) | 1446 (2.9) | 115 (2.7) | 72 (3.0) | 38 (3.1) |
| No | 52026 (90.9) | 44874 (91.0) | 3892 (90.8) | 2154 (88.8) | 1106 (88.8) |
| Missing | 3557 (6.2) | 2975 (6.0) | 281 (6.6) | 200 (8.2) | 101 (8.1) |
| Cesarean delivery, n (%) | 18767 (32.8) | 14669 (29.8) | 2059 (48.0) | 1237 (51.0) | 802 (64.4) |
| COPS2 score at admission, mean (SD) | 10.10 (3.39) | 10.06 (3.16) | 10.30 (4.34) | 10.28 (4.43) | 10.53 (5.50) |
| Preterm delivery, n (%) | 7308 (12.8) | 4248 (8.6) | 1351 (31.5) | 917 (37.8) | 792 (63.6) |
Among patients discharged on antihypertensive medication, most were prescribed either labetalol only (54%) or nifedipine only (30%). Demographic and clinical characteristics of patients prescribed labetalol only were similar to those prescribed nifedipine only. However, patients prescribed two or more antihypertensive medications were more likely than patients prescribed other antihypertensive medication to have severe preeclampsia (71% vs 49–57%), to be Black (20% vs 12%), to have had a Cesarean delivery (64% vs 48–51%), and to have delivered preterm (64% vs 32–38%) (Table 2).
Unadjusted cumulative incidence of readmission varied by antihypertensive medication prescribed (Figure 1). By 3 days after discharge, 0.8% of patients discharged without medication had been readmitted, while 1.9% of patients discharged on nifedipine only, 3.8% of patients discharged on labetalol only, and 1.4% discharged on 2 or more medications had been readmitted. Similarly, by 2 weeks post-discharge, 2.1% of patients discharged without medication had been readmitted, while 3.1% of patients discharged on nifedipine only, 7.2% of patients discharged on labetalol only, and 2.6% of patients discharged on 2 or more medications had been readmitted.
Figure 1.
Unadjusted cumulative incidence curves showing probability of postpartum readmission within 6 weeks for each type of medication at discharge (no antihypertensive medication; labetalol only; nifedipine only; multiple medications) among 57,254 HDP pregnancies.
In adjusted statistical models, type of antihypertensive medication prescription at discharge was strongly associated with incidence of readmission. Compared to patients discharged without antihypertensive medication, discharge on labetalol only was associated with 63% (HR: 1.63; 95%CI: 1.41, 1.88) greater incidence of postpartum readmission in adjusted models. In contrast, discharge on nifedipine only and discharge on multiple antihypertensive medications were associated with 26% and 47% lower incidence of postpartum readmission, respectively ((HR: 0.74 (95% CI: 0.59, 0.93) and 0.53 (95% CI: 0.38, 0.74), in adjusted models) (Table 3). After excluding patients with pre-pregnancy hypertension, we still found elevated incidence of readmission among patients prescribed labetalol only and reduced incidence among patients prescribed two or more antihypertensive medications (versus no medication), though nifedipine was no longer associated with reduced incidence of readmission (Supplemental Table S1). In models with labetalol monotherapy as the reference group, nifedipine and two or more antihypertensive medications were associated with 50% and 62% lower incidence of readmission, respectively (Supplemental Table S2).
Table 3.
Associations between antihypertensive medication prescription at discharge with incidence of readmission within 6 weeks
| Antihypertensive medication at discharge | n | Percent readmitted | Unadjusted hazard ratio (95%CI) | Adjusted hazard ratio (95%CI) |
|---|---|---|---|---|
| No antihypertensive medication | 49295 | 2.5 | Reference | Reference |
| Labetalol only | 4288 | 7.6 | 3.14 (2.78, 3.55) | 1.63 (1.41, 1.88) |
| Nifedipine only | 2426 | 3.6 | 1.45 (1.17, 1.80) | 0.74 (0.59, 0.93) |
| Two or more antihypertensive medications | 1245 | 3.2 | 1.28 (0.94, 1.76) | 0.53 (0.38, 0.74) |
Models adjusted for hypertensive diagnosis, last systolic and diastolic blood pressure, age, pre-pregnancy BMI, mode of delivery, insurance type, race/ethnicity, facility, smoking, preterm delivery, parity, neighborhood deprivation index, delivery year, and COPS2 score at admission.
There was no evidence that the effect of discharge antihypertensive medication type on incidence of readmission varied by race and ethnicity (interaction p=0.88). Specifically, among patients discharged without antihypertensive medication, Black patients had 60% greater incidence of readmission compared to White patients. This elevated incidence of readmission was consistent among patients discharged on labetalol only (i.e., Black patients had 49% greater incidence of readmission than White patients) and on nifedipine only (i.e., Black patients had 52% greater incidence of readmission compared to White patients), though confidence intervals were wide (Figure 2).
Figure 2.
Incidence of 6-week postpartum readmission for patients of each racial and ethnic group within categories of discharge antihypertensive medication prescription, compared to a reference group of White patients. Effect estimates represent a race/ethnicity*medication product term (interaction p = 0.88) from models adjusted for hypertensive diagnosis, last systolic and diastolic blood pressure, age, pre-pregnancy BMI, mode of delivery, insurance type, facility, smoking, preterm delivery, parity, neighborhood deprivation index, delivery year, and COPS2 score at admission.
Comment
Principal Findings
Among patients with HDP in a large and racially diverse integrated health care delivery system, we found that postpartum discharge on labetalol only (versus no medication) was associated with greater incidence of readmission within 6 weeks, even after adjustment for clinical and demographic factors including last inpatient blood pressure and type of HDP. Additionally, discharge on nifedipine alone or multiple antihypertensive medications versus no medication were associated with lower incidence of postpartum readmission. Findings suggest varying effectiveness among discharge medications at preventing postpartum readmission in a population of patients with HDP, who are at elevated risk of postpartum readmission.
Results in the Context of What is Known
Our finding that discharge on labetalol alone was associated with increased incidence of postpartum readmission is consistent with two previous observational studies. In an analysis of claims data, Do and colleagues reported that postpartum patients discharged on labetalol alone had 63% higher odds of readmission compared to patients discharged on nifedipine alone.15 Similarly, in a retrospective cohort from a tertiary care center serving a mostly White patient population, Lovgren and colleagues reported that patients with peripartum hypertension discharged on labetalol alone had 66% higher odds of readmission and those discharged on nifedipine alone had 73% lower odds of readmission compared to patients discharged without antihypertensive medication.16 Both studies also reported that patients discharged on both nifedipine and labetalol had lower odds of readmission (20% lower compared to nifedipine alone15 and 65% lower compared to no antihypertensive medication,16 respectively). Importantly, our results were consistent with Do et al.’s findings even after adjusting for type of HDP, last inpatient blood pressure, parity, and pre-pregnancy BMI, predictors of readmission which were not available in their dataset. Our analysis in a large, racially diverse cohort with excellent confounding control therefore confirms higher readmission incidence after discharge on labetalol alone, and lower readmission incidence after discharge on nifedipine alone, versus no antihypertensive medication at discharge after HDP.
There are several potential explanations for reduced effectiveness of labetalol to prevent postpartum readmission. First, nifedipine is thought to improve renal blood flow,29 potentially leading to quicker resolution of transient postpartum hypervolemia compared to treatment with labetalol, which acts through peripheral vasodilation.30 Alternatively, post-discharge medication adherence may be easier with nifedipine than labetalol: nifedipine is available in an extended-release formulation, allowing patients to take it as infrequently as once per day, while labetalol must be taken at least twice a day.15 Potential mechanisms deserve additional investigation.
Research Implications
Our finding of heterogeneous effectiveness among discharge antihypertensive medications to prevent readmission may explain why prior studies have found mixed results when evaluating the effect of discharge medications as a group (i.e., testing the effectiveness of any versus no antihypertensive medication).5,10 Clinical evidence permits use of several classes of antihypertensive medication for postpartum blood pressure management,31 and labetalol is commonly prescribed: approximately half of HDP patients in our cohort discharged on antihypertensive medication received a prescription for labetalol alone. However, comparative effectiveness of oral labetalol versus oral nifedipine for postpartum blood pressure management has not been evaluated outside of small (n<100 per arm) clinical trials.32,33 Additionally, the trials focused on short-term postpartum blood pressure control rather than readmission risk, which may be a function of both medication effectiveness and adherence after discharge. Our results increase the urgency of formally evaluating the effectiveness of specific discharge medication regimens among the high-risk postpartum HDP patient population.
Clinical Implications
Racial inequities in readmission risk have been well documented9 and were present in our study. Lower risk of readmission after discharge on nifedipine and multiple antihypertensive medications versus no medication on average may indicate that increasing use of nifedipine or multiple antihypertensive medications among Black patients could reduce readmission risk. On the other hand, we observed consistently elevated rates of readmission for Black versus White patients among those discharged without medication, with labetalol monotherapy, and with nifedipine monotherapy, which may suggest that medication choice may not substantially impact disparities in readmission. Given the variation in readmission risk among antihypertensive medication regimens, future research should examine the extent to which altered discharge antihypertensive medication prescribing could impact racial and ethnic disparities in postpartum readmission.
Strengths and Limitations
Strengths of our analysis include use of KPNC’s EHR, which provides detailed information on nearly 60,000 HDP-complicated pregnancies about inpatient blood pressures, hypertensive diagnoses, and demographic information including pre-pregnancy BMI, variables rarely included in large obstetric databases.3,34 Similarly, we leveraged the EHR to collect discharge medication information, allowing us to compare risks of readmission among specific prescribed medications and combinations of medications. Finally, because KPNC members typically receive all medical care at KPNC facilities, we were likely able to fully capture postpartum readmissions.
Our analysis also had limitations. We did not analyze reasons for patients’ postpartum readmissions, so patients with greater healthcare engagement may have been more likely to be screened with a postpartum blood pressure measurement and readmitted. However, readmission indicates severe health concerns, so associations are unlikely to be mostly explained by increased screening. We did not ascertain indication for readmission, but readmissions for indications unrelated to hypertension would be expected to occur equally frequently in all discharge medication groups and are unlikely to introduce a spurious association. It is possible that some patients initiated antihypertensive medications after discharge rather than receiving a discharge prescription, or did not take prescribed discharge medications, which we did not assess; we also did not collect details about medication dosage. However, we still reported strong findings, which may reflect the real-world effectiveness of discharge antihypertensive prescription. We did not collect information about how providers chose antihypertensive medications, but the demographics of patients prescribed labetalol monotherapy versus nifedipine monotherapy were similar. Multiple correlated clinical factors have been shown to strongly predict risk of readmission; we were able to adjust for many important predictors but cannot rule out the possibility of residual confounding. Diagnostic criteria for preeclampsia changed in 2013, which may have introduced measurement error into our assessment of HDP type; however, minor misclassification in this adjustment covariate is unlikely to substantially affect results. Finally, some patients with pre-existing hypertension may resume their pre-pregnancy antihypertensive regimen after delivery, leading to lower readmission risk in this group; however, labetalol was associated with greater readmission risk even after removing patients with pre-pregnancy hypertension.
Conclusions
In this study of pregnancies with HDP, we found that discharge on labetalol only (versus no medication) was associated with greater incidence of postpartum readmission, while discharge on nifedipine alone or multiple antihypertensive medications were associated with lower incidence of postpartum readmission. Investigation into comparative effectiveness of commonly prescribed discharge medications for the management of postpartum hypertension after HDP is warranted.
Supplementary Material
AJOG at a Glance:
A. Why was this study conducted?
Patients with hypertensive disorders of pregnancy have high risk of postpartum readmission.
Choice of antihypertensive medication may represent an actionable clinical target to reduce postpartum readmissions.
B. What are the key findings?
Compared to discharge without antihypertensive medication, discharge on labetalol only was associated with 63% greater incidence of postpartum readmission.
Discharge on nifedipine only or on two or more antihypertensive medications were associated with 26% and 47% lower incidence of postpartum readmission, respectively.
C. What does this study add to what is already known?
In a large and diverse population of patients with excellent confounding control, we confirm previous preliminary findings of higher readmission risk after discharge on labetalol alone, and lower readmission risk after discharge on nifedipine alone.
Financial support:
Research reported in this publication was supported by the Kaiser Foundation Health Plan, The Permanente Medical Group, and the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health under Award Number K12AR084219. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Disclosure statement: The authors report no conflict of interest.
References
- 1.Liu S, Heaman M, Joseph KS, et al. Risk of maternal postpartum readmission associated with mode of delivery. Obstetrics & Gynecology. 2005;105(4):836–842. [DOI] [PubMed] [Google Scholar]
- 2.Clapp MA, Little SE, Zheng JS, Robinson JN. A multi-state analysis of postpartum readmissions in the United States. Am J Obstet Gynecol. 2016;215(1):113.e111–113.e110. [DOI] [PubMed] [Google Scholar]
- 3.Mogos MF, Salemi JL, Spooner KK, McFarlin BL, Salihu HH. Hypertensive disorders of pregnancy and postpartum readmission in the United States: national surveillance of the revolving door. J Hypertens. 2018;36(3):608–618. [DOI] [PubMed] [Google Scholar]
- 4.Emeruwa UN, Gyamfi-Bannerman C, Wen T, et al. Adverse outcomes during postpartum readmissions after deliveries complicated by hypertensive disorders of pregnancy. Am J Perinatol. 2022;39:699–706. [DOI] [PubMed] [Google Scholar]
- 5.Bruce KH, Anderson M, Stark JD. Factors associated with postpartum readmission for hypertensive disorders of pregnancy. Am J Obstet Gynecol. 2021;3:100397. [DOI] [PubMed] [Google Scholar]
- 6.McLaren RA, Magenta M, Gilroy L, et al. Predictors of readmission for postpartum preeclampsia. Hypertens Pregnancy. 2021;40(3):254–260. [DOI] [PubMed] [Google Scholar]
- 7.DeVries CR, Starbird WT, Arab A, Bailey B. Investigation of factors that predict risk for hospital readmission following delivery of pregnancies complicated by hypertensive disorders of pregnancy. J Maternal Fetal Neonatal Med. 2022;35(25):9320–9324. [DOI] [PubMed] [Google Scholar]
- 8.Stamilio DM, Beckham J, Boggess KA, Jelovsek JE, Venkatesh KK. Risk factors for postpartum readmission for preeclampsia or hypertension before delivery discharge among low-risk women: a case-control study. Am J Obstet Gynecol. 2021;3(100317). [DOI] [PubMed] [Google Scholar]
- 9.Wagner JL, White RS, Tangel V, Gupta S, Pick JS. Socioeconomic, racial, and ethnic disparities in postpartum readmissions in patients with preeclampsia: a multi-state analysis, 2007–2014. J Racial Ethn Health Disparities. 2019;6:806–820. [DOI] [PubMed] [Google Scholar]
- 10.Lovgren T, Connealy B, Yao R, Dahlke JD. Postpartum management of hypertension and effect on readmission rates. Am J Obstet Gynecol. 2022;4:100517. [DOI] [PubMed] [Google Scholar]
- 11.American College of Obstetricians and Gynecologists. Executive summary: Hypertension in Pregnancy. Obstetrics & Gynecology. 2013;122(5):1122–1131. [DOI] [PubMed] [Google Scholar]
- 12.Hoffman MK, Ma N, Roberts A. A machine learning algorithm for predicting maternal readmission for hypertensive disorders of pregnancy. Am J Obstet Gynecol MFM. 2021;3:100250. [DOI] [PubMed] [Google Scholar]
- 13.Mukhtarova N, Alagoz O, Chen YH, Hoppe K. Evaluation of different blood pressure assessment strategies and cutoff values to predict postpartum hypertension-related readmissions: a retrospective cohort study. Am J Obstet Gynecol. 2021;3:100252. [DOI] [PubMed] [Google Scholar]
- 14.Balhotra K, Roach C, Al-Kouatly HB, Minkoff H, McLaren RA Jr. Association of antihypertensive medication at discharge with readmission for postpartum preeclampsia. American Journal of Obstetrics & Gynecology. 2023;In Press. [DOI] [PubMed] [Google Scholar]
- 15.Do SC, Leonard SA, Kan P, et al. Postpartum readmission for hypertension after discharge on labetalol or nifedipine. Obstet Gynecol. 2022;140:591–598. [DOI] [PubMed] [Google Scholar]
- 16.Lovgren T, Connealy B, Yao R, Dahlke JD. Postpartum medical management of hypertension and risk of readmission for hypertensive complications. Journal of Hypertension. 2023(41):351–355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Ghosh G, Grewal J, Männistö T, et al. Racial/ethnic differences in pregnancy-related hypertensive disease in nulliparous women. Ethn Dis. 2014;24(3):283–289. [PMC free article] [PubMed] [Google Scholar]
- 18.Gordon NP. Similarity of the Adult Kaiser Permanente Membership in Northern California to the Insured and General Population in Northern California: Statistics from the 2011–12 California Health Interview Survey. Oakland, CA: Kaiser Permanente Division of Reseach;2013. [Google Scholar]
- 19.American College of Obstetricians and Gynecologists. Gestational hypertension and preeclampsia. ACOG Practice Bulletin No. 222. Obstetrics & Gynecology. 2020;135(6):e237–e260. [DOI] [PubMed] [Google Scholar]
- 20.Armstrong MA, Lieberman L, Carpenter DM, et al. Early Start: an obstetric clinic-based, perinatal substance abuse intervention program. QualManagHealth Care. 2001;9(2):6–15. [DOI] [PubMed] [Google Scholar]
- 21.World Health Organization. The International Classification of adult underweight, overweight and obesity according to BMI 2012.
- 22.American College of Obstetricians and Gynecologists. Methods for Estimating the Due Date. Committee Opinion No. 700. Obstet Gynecol. 2017;129(5):e150–e154. [DOI] [PubMed] [Google Scholar]
- 23.Menard MK, Main EK, Currigan SM. Executive summary of the reVITALize initiative: standardizing obstetric data definitions. Obstet Gynecol. 2014;124(1):150–153. [DOI] [PubMed] [Google Scholar]
- 24.Hedderson MM, Ferrara A, Sacks DA. Gestational diabetes mellitus and lesser degrees of pregnancy hyperglycemia: association with increased risk of spontaneous preterm birth. Obstet Gynecol. 2003;102(4):850–856. [DOI] [PubMed] [Google Scholar]
- 25.Messer LC, Laraia BA, Kaufman JS, et al. The development of a standardized neighborhood deprivation index. J Urban Health. 2006;83:1041–1062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–383. [DOI] [PubMed] [Google Scholar]
- 27.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613–619. [DOI] [PubMed] [Google Scholar]
- 28.Stoddard PJ, Laraia BA, Warton EM, et al. Neighborhood deprivation and change in BMI among adults with type 2 diabetes: the Diabetes Study of Northern California (DISTANCE). Diabetes Care. 2013;36(5):1200–1208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Barton JR, Hiett AK, Conover WB. The use of nifedipine during the postpartum period in patients with severe preeclampsia. Am J Obstet Gynecol. 1990;162(3):788–792. [DOI] [PubMed] [Google Scholar]
- 30.Chandiramani M, Shennan A, Waugh J. Modern management of postpartum hypertension. Trends in Urology Gynaecology & Sexual Health. 2007;12(3):37–42. [Google Scholar]
- 31.Kumar NR, Hirschberg A, Srinivas SK. Best practices for managing postpartum hypertension. Current Obstetrics and Gynecology Reports. 2022;11:159–168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Sharma KJ, Greene N, Kilpatrick SJ. Oral labetalol compared to oral nifedipine for postpartum hypertension: A randomized controlled trial. Hypertens Pregnancy. 2017;36(1):44–47. [DOI] [PubMed] [Google Scholar]
- 33.Ainuddin JA, Javed F, Kazi S. Oral labetalol versus oral nifedipine for the management of postpartum hypertension a randomized controlled trial. Pak J Med Sci. 2019;35(5):1428–1433. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Wen T, Yu VX, Wright JD, et al. Postpartum length of stay and risk for readmission among women with preeclampsia. J Maternal Fetal Neonatal Med. 2020;33(7):1086–1094. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.