Keywords
infective endocarditis - pregnancy - intravenous drug use - bacteremia - opioid use
disorder
Opioid use disorder (OUD) and its associated complications in the reproductive age
population are reaching epidemic proportions.[1]
[2] As rates of OUD continue to rise, the incidence of pregnancy-related sequalae of
OUD such as peripartum maternal opioid overdoses, neonatal abstinence syndrome, and
hepatitis C infections has also increased.[3-9] The incidence of infective endocarditis (IE), a complication of OUD most commonly
associated with intravenous drug use (IDU), is also increasing.[10]
[11] Due to a historically low incidence, literature on IE in pregnancy is mostly limited
to individual case reports and systematic reviews.[12] Although treatment of IE in pregnancy is extrapolated from general population research,
there are added complexities of treating IE associated with IDU in the antenatal or
perinatal periods that require special attention.[13]
[14]
[15] In addition to the acute treatment of IE, treatment for addiction, support for chronic
psychosocial conditions, and knowledge of state-specific medicolegal environments
may be needed to optimize outcomes. Given the current opioid epidemic, we noted an
increase in pregnancy-associated admissions for IE stemming from IDU at our institution.
As an institution with a dedicated Substance Abuse Treatment and Prevention Program
for opioid dependent pregnancies, our objective is to add to the existing literature
on the management of IE resulting from IDU in pregnancy and highlight important aspects
of antepartum care for IE as they relate to OUD and IDU.
Methods
In this retrospective cohort case series, we identified women with a diagnosis of
IE related to IDU during a pregnancy episode from years 2012 to 2019. A search of
electronic medical records for pregnancy episodes with the associated diagnosis of
IE was conducted. International Classification of Diseases-9th revision/10th revision
(ICD-9/ICD-10) diagnostic codes for diseases of the circulatory system and IE complicating
pregnancy were used for the search. In addition, combinations of keywords such as
“infective endocarditis,” “opioid use disorder,” and “polysubstance abuse” were used
to search the Epic Integrated Health Information Systems patient “Problem List.” The
“Problem List” tracks active and historical diagnoses entered into the patient's charts
by providers. A total of 57 patients were identified. This initial cohort was screened
for IE during an active pregnancy episode by using Duke's modified criteria, namely
the presence of positive blood cultures and echocardiographic findings consistent
with endocardial vegetations. Women who were not persons who inject drugs or who did
not have acute IE during a pregnancy episode (antepartum or within 4 weeks postpartum)
were excluded ([Fig. 1]—create flow diagram)
Fig. 1 Patient selection flowchart. IDU, intravenous drug use; IE, infective endocarditis.
The medical records of the 12 women who met inclusion criteria were reviewed. Data
on demographics, past medical and obstetrical histories, drug use history, details
of antepartum hospital admissions and treatment courses for IE, echocardiographic
data, initiation of opioid agonist therapy or medication-assisted treatment, prenatal
outpatient follow-up, pregnancy outcomes including delivery complications and neonatal
outcomes, postpartum follow-up, and hospital readmissions in the peripartum period
were collected. Descriptive statistics (mean ± standard deviation for normally distributed
continuous variables and median and interquartile range for non-normally distributed
continuous variables; frequencies and percentages for categorical variables) were
calculated. This study was approved by The Ohio State University Institutional Review
Board.
Results
Patient Characteristics
The 12 women within our cohort had a mean age of 30 ( ± 4.58). The majority was white
(80%) ([Table 1]). All were known to be persons who inject drugs and the majority (67%) received
prenatal care through our substance abuse specialty prenatal clinic. The vast majority
had a coexisting diagnosis of hepatitis C (92%). Nine women were treated for endocarditis
in the antepartum period and two within 4 weeks of the postpartum period. One patient
presented for new obstetric care in the setting of incompletely treated endocarditis
prior to conception. One patient was discharged without antibiotic therapy after additional
imaging proved her cardiac lesion to be more consistent with previously treated IE.
The majority of women (84%) presented with cardiopulmonary symptoms including shortness
of breath, dyspnea on exertion, chest pain with or without fever. The remainder presented
with symptoms related to the infectious sequalae of IDU, including extremity pain
from injection site, abscess, and hip pain secondary to sacroiliitis (16%).
Table 1
Demographic and clinical characteristics of patients with IE
|
Characteristics
|
|
|
Age (y), mean (± SD)
|
30.2 ± 4.58
|
|
Gestational age at diagnosis (wk), mean (IQR)
|
25.4 (21–31)
|
|
Caucasian
|
|
Antepartum diagnoses (n)
|
10 (83%)
|
|
Postpartum diagnoses (n)
|
2 (17%)
|
|
Comorbidities
|
n (%)
|
|
Hepatitis C
|
11 (92)
|
|
Psychiatric disorder[a]
|
3 (25)
|
|
Valve involvement
|
n (%)
|
|
Native valve
|
11 (92)
|
|
Prosthetic
|
1 (8)
|
|
Left sided, single
|
2 (17)
|
|
Mitral
|
2 (100)
|
|
Aortic
|
0 (0)
|
|
Right sided, single
|
8 (66)
|
|
Tricuspid
|
7 (86)
|
|
Pulmonic
|
1 (14)
|
|
Multivalvular (both sides)
|
2 (17)
|
Abbreviations: IE, infective endocarditis; IQR, interquartile range; SD, standard
deviation.
a Psychiatric disorders including depression, anxiety, bipolar disorder.
Echocardiographic Findings
Echocardiograph results were available for all women in our cohort ([Table 2]). The majority of these studies was transthoracic versus transesophageal echocardiograms
(91%). Discrete vegetations were visualized in 80% of the patients, mostly commonly
involving the tricuspid valve (56%). When measured quantitatively, valvular lesions
ranged from 0.6 × 0.8 cm to 3.0 × 1.6 cm in size. Otherwise, lesions were classified
as either small or large in echocardiographic reports. Interval echocardiographs were
available for six women and the majority of these (83%) had interval decrease in vegetation
size after antibiotic treatment.
Table 2
Microbiologic and treatment data
|
n (%)
|
|
Bacteria
|
|
Methicillin-resistant Staphylococcus aureus
|
5 (42)
|
|
Methicillin-sensitive Staphylococcus aureus
|
4 (33)
|
|
Beta-hemolytic streptococcus, Group C
|
1 (8)
|
|
None
|
2 (17)
|
|
Treatment
|
|
Antibiotic regimen complete
|
6 (60)
|
|
Length of hospital stay (median, IQR)
|
26 (16–39)
|
|
Resolution of vegetation
|
4 (66)
|
Abbreviation: IQR, interquartile range.
Blood Cultures and Antibiotic Therapy
Bacteremia with Staphylococcal aureus was identified in three-fourths of the women (75%). Vancomycin was the most commonly
used antibiotic (58%); however, transition from vancomycin due to side effects occurred
in five cases. Daptomycin was the most frequently used alternative, but ceftaroline
fosamil and linezolid were also used. Recommended antibiotic course lengths ranged
from 2 to 8 weeks. Approximately half of these women did not complete their recommended
antibiotic course due to discharge against medical advice (AMA). Circumstances surrounding
discharge AMA were not consistently specified in the medical records. Reasons cited
included continued illicit drug use while inpatient (e.g., crushing oral opioids for
self-injection) and the reported need to handle family affairs outside of the hospital
setting.
Opioid Use Disorder
All women were known to be persons who inject drugs and the majority (67%) received
prenatal care through our substance abuse specialty prenatal clinic. Urine toxicology
report revealed that all women had tested positive for opioids either upon admission
or in the recent past preceding current admission. Other nonprescribed, illicit substances
discovered in toxicology reports included cocaine, amphetamines, tetrahydrocannabinol,
and benzodiazepines. During the time of admission, the majority of women (92%) received
treatment for OUD through either the initiation or continuation of opioid agonist
and antagonist medications such as suboxone (67%), methadone (17%), or Subutex (1%)
([Table 3]). Inpatient psychosocial interventions were routinely offered to all women in conjunction
with psychiatric and social work consultation services. Additionally, all women had
the option of continuing prenatal care through our substance abuse specialty prenatal
clinic where peer support is available through group prenatal care.
Table 3
Opioid use disorder data
|
n (%)
|
|
Urine toxicology results
|
|
Opioids (heroin or prescription)
|
12 (100)
|
|
Cocaine
|
5 (42)
|
|
Amphetamines
|
3 (25)
|
|
THC
|
3 (25)
|
|
Benzodiazepines
|
3 (25)
|
|
Treatment of OUD during admission
|
|
Suboxone
|
8 (67)
|
|
Methadone
|
2 (17)
|
|
Subutex
|
1 (8)
|
|
Neonatal abstinence syndrome
|
|
NICU admission
|
4 (44)
|
Abbreviations: NICU, neonatal intensive care unit; THC, tetrahydrocannabinol; OUD,
opioid use disorder.
Outcomes
Median gestational age at diagnosis was 25 weeks (interquartile range [IQR] 21–31)
and median gestational age at delivery was 37 weeks (35–38). There were four cesarean
deliveries (33%) for usual obstetrical indications. The median duration of hospital
course was 26 days (16–39). Besides OUD and hepatitis C virus, septic pulmonary emboli
(73%), anemia of chronic disease (55%), and acute kidney injury (45%) were the most
commonly encountered comorbidities. There was one maternal death, after urgent cardiac
surgery for heart failure secondary to severe aortic and tricuspid valve vegetations
and ventricular septum abscess. The remaining patients had clinical improvement or
interval improvement of echocardiographic findings. Sixty-seven percent of patients
were initiated on medication-assisted treatment during their hospitalization and were
continued on treatment in our OUD program.
Over half of pregnancies delivered at term (55%) and the majority of preterm deliveries
was iatrogenic (75%) for reasons unrelated to OUD or IE. The one case of maternal
cardiac surgery that required extracorporeal membrane oxygenation resulted in intrauterine
fetal demise at 19 weeks gestation. There was one case of a term fetal demise at 38
weeks gestation as a result of placental abruption. Four of the neonates (44%) required
neonatal intensive care unit admission and pharmacologic treatment for neonatal abstinence
syndrome of varying severity. All living newborns were discharged between day of life
4 and 25 days (median = 12, [5–16]).
The majority of patients was lost to cardiothoracic surgery follow-up (64%); however,
none presented for recurrent IE within our 4 weeks follow-up period. Thus, a third
did have subsequent clinical encounters stemming from continued IDU (33%).
Discussion
IE during pregnancy is rare with a reported incidence of 0.006%.[13]
[16] A recent systematic review of peripartum IE reported maternal and fetal mortality
rates of 11 and 14%; these rates have traditionally been higher, up to 33 and 29%
for mother and fetus, respectively.[13]
[17]
[18] The pathogenic mechanism of IE in IDU involves endothelial damage to cardiac valves
from intravascular injection of foreign particulate matter while using illicit drugs.[19] In a similar manner, bacterial contaminant from skin or contaminated syringes may
also be inadvertently injected into the blood stream during drug usage.[20]
IE should be suspected in any patient presenting with a fever and a new onset heart
murmur; the clinical and echocardiographic features of Duke's criteria can be used
to formally make the diagnosis. Major diagnostic criteria include positive blood cultures
for microorganisms typical of IE and echocardiographic evidence of endocardial involvement.[13] Alternatively, septic pulmonary emboli, pneumonia, or positive blood cultures for
common skin contaminants in the context of IDU history may prompt echocardiographic
imaging and subsequent diagnoses of IE.[15]
The most commonly identified bacterial pathogens in patients with IE are Staphylococcus aureus and Streptococcus viridians.[21]
[22] Courses of up to 8 weeks of intravenous antibiotics may be indicated for methicillin-resistant
Staphylococcus aureus, whereas shorter courses may suffice for uncomplicated IE due to isolated methicillin-sensitive
Staphylococcal bacteremia.[23] For cases of acute heart failure or those refractory to medical management, cardiac
surgery for valve replacement may be necessary. Cardiac surgery and cardiopulmonary
bypass carry significant risk of mortality, up to 30 and 6% for fetus and mother,
respectively.[24]
[25] In pregnant women, with nonurgent indications for surgery, the fetal benefits of
an advancing gestation must be weighed against the potential of an acutely worsening
cardiac status. Literature for cardiac surgery and cardiopulmonary bypass in pregnancy
suggests elective delivery prior to maternal cardiac surgery improves fetal outcome
at later gestational ages.[26] In theory, high-flow, high-pulsatile, normothermic CPB may improve uterine blood
flow although clinical studies for optimal settings are lacking.[27]
[28]
Successful antepartum care of patients with IE in pregnancy may be fraught with difficulties
stemming from concomitant OUDs, the nature of extended-length hospital stays for the
treatment of IE, AMA discharges, and chronic comorbidities.[14]
[29] Among persons who use drugs, IDU, in particular, has been implicated as an independent
risk factor for discharge AMA.[30]
[31] Previous research highlights the negative impact discharge AMA may have for pregnancies
in vulnerable patient populations such as women with substance abuse.[32] In contrast, in-hospital opioid agonist therapy has been demonstrated as a negative
risk factor for discharge AMA in this same population. In addition to in-hospital
retention, the potential benefits of initiating opioid agonist therapy for hospitalized
antepartum patients include improved obstetrical outcomes that are related to the
discontinued use of illicit substances and improved prenatal care compliance.[33]
The historical barriers to successful initiation and maintenance of medication-assisted
treatment for OUD are manifold. In pregnancy, a lack of screening and under diagnosis
predisposed to suboptimal rates of treatment of OUD in antepartum period.[34]
[35] Early research controversies suggesting an association between opioid exposure and
birth defects, along with concerns regarding risks of neonatal abstinence syndrome,
created reservations for the routine prescription of opioid agonists to treat OUD.[36]
[37] Lastly, when treatment for OUD was actually pursued, approaches associated with
failure (e.g., lack of third trimester up-titration) or high rates of relapse (e.g.,
medically supervised withdrawal) may have been deployed.[38]
[39] Time and research have addressed many of these controversies; however, large proportions
of pregnant women still do not receive opioid agonist therapies when hospitalized.[40]
[41]
[42]
[43]
[44]
State-mandated reporting and the threat of civil or criminal penalties including the
loss of child custody hinder the physician–patient therapeutic alliance and potentially
discourage women from seeking and completing needed therapy for sequelae of OUD such
as intravenous drug user (IVDU)-associated IE.[45]
[46]
[47] As of May 2020, 25 states require healthcare providers to report suspected prenatal
drug use and 23 states consider such child abuse under civil child-welfare statutes,
whereas only 19 states have targeted drug treatment programs for pregnant women.[48] In addition to advocating for less punitive legislation for drug use in pregnancy,
treating obstetrician–gynecologists should be knowledgeable of their individual state's
policies and be wary of the role such policies may have in the nonadherence to treatment
regimens.
In addition to addressing OUD, healthcare professionals should be aware of the other
chronic comorbid conditions that may complicate care during antepartum hospitalizations.
Women with OUD have a high prevalence of undiagnosed, chronic mood disorders. Fitzsimons
prospectively observed a greater than 70% rate of co-occurring mood or anxiety disorders
for pregnant women entering into a treatment center for substance use disorders.[49] At equipped hospital facilities, women admitted for the treatment of IVDU-associated
IE should be screened for and have management of underlying psychosocial conditions
in consultation with appropriate psychiatric and social work services.
Even with optimal management of underlying addiction and comorbid psychiatric conditions,
the extended courses needed to complete therapy for IE in the hospital setting may
be difficult for any woman. Strategies of motivational interviewing, as well as the
utilization of shorter duration treatment courses for uncomplicated IE may improve
adherence. Restrictive hospital policies during these admissions (e.g., off-ward privileges,
limited guess access), aimed at reducing the risks of continued, surreptitious drug
use, may inadvertently increase rates of discharge AMA. Ultimately, an increase in
the availability of long-term acute health facilities that accept pregnant women may
be needed. Recent data highlights the importance of women-centered treatment programs
in improving pregnancy-specific outcomes such as postpartum visit attendance, breastfeeding
continuation rates, and long-acting contraceptive utilization.[50]
[51]
Conclusion
IE in pregnancy, as it relates to IDU, may be associated with significant maternal
and fetal morbidity. In pregnant patients with OUD and IE, additional considerations
must be taken into account to improve outcomes. The initiation of opioid agonist therapy,
psychosocial support, and services to address underlying psychiatric diagnoses, improved
access to long-term acute health facilities that accept pregnant women, and strategies
to minimize AMA discharge and incomplete treatment must all be considered to optimize
treatment of this vulnerable population.
