Pulmonary embolism is associated with current morphine treatment in patients with deep vein thrombosis.

Pulmonary embolism is associated with current morphine

treatment in patients with deep vein thrombosis

Cynthia Wei-Sheng Lee1, Chih-Hsin Muo2, Ji-An Liang3,4, Fung-Chang Sung2,3,

Chia-Hung Kao3,5 and

Jun-Jun Yeh6,7,8

Introduction

Pulmonary embolism (PE), caused by venous thromboembolism

(VTE) from the lower extremities or

locally formed pulmonary artery thrombosis (1), is a

relatively common vascular disease and is the underlying

cause of a significant number of sudden deaths (2,

3). In the United States, the incidence rate exceeds 1 per

1000 patients, and the mortality rate is >15% in the

first 3 months after diagnosis (4), showing that PE is

nearly as deadly as acutemyocardial infarction (5). The

risk factors of PE include prolonged immobilization,

surgery, thrombophlebitis, bone fracture of lower limb,

estrogen use and pregnancy/postpartum (6). PE is

usually a consequence of deep vein thrombosis (DVT)

(7) because the venous clot may dislodge and migrate

to the lungs.

Illicit opioid users are at increased risk of hospital

admission for DVT, and aging, being a woman and

intravenous delivery all independently increase the risk

of developing DVT (8). Morphine has been found to

antagonize prostaglandin E1-mediated inhibition of

human platelet aggregation (9). Furthermore, morphine

may exert its potentiation in platelet aggregation

by binding to α2-adrenoceptors in human platelets,

thereby reducing intracellular cAMP formation and

increasing the activation of phospholipase C and the

Na+/H+ exchanger (10). Therefore, morphine might

promote the formation of thrombosis by promoting

platelet aggregation, resulting in DVT and the ensuing

PE.

often administered morphine, it is important to understand

the relationship between morphine use and PE

incidence. No epidemiological study has investigated

this relationship in Taiwan. To evaluate the potential

for morphine-induced PE, we compared the incidence

of PE in DVT patients treated with and without morphine

using data from the National Health Insurance

Research Database (NHIRD) of Taiwan.

Materials and methods

For this nested case-control study, we used Taiwan

National Health Insurance (NHI) claims records that

were accessible in the NHIRD. The universal singlepayer

health-care program was instituted on March 1,

1995, and reached a coverage rate of over 99% of Taiwanese

residents by 2010. This database contains the

data of one million people covered by the Taiwan NHI

program; each patient was randomly selected from all

beneficiaries in 2000, and includes all outpatient and

inpatient records from 1996 to 2010.We confirm that

all data was de-identified and analyzed anonymously.

In addition, this study was also approved by the Ethics

Review Board at China Medical University

(CMUREC-101-012).

We obtained the data of 3668 patients with newly

diagnosed DVT [International Classification of Diseases,

9th Revision, ClinicalModification (ICD-9-CM)

453.8] for the period 1998–2010 to serve as the DVT

cohort for this study. We excluded 55 patients with a

history of PE (ICD-9-CM 415.1) prior to the date of

their DVT diagnosis. All 174 DVT patients who developed

PE before the end of 2010 were designated as the

case group, and the date of PE diagnosis was defined as

the index date. Eight controls were selected from DVT

patients who did not develop PE before the end of 2010

as the control group based on the DVT year and index

year of each case.

Variables included age, sex, morphine use and

comorbidities. Among the comorbidities specified

were lower limb injuries (including fracture and

surgery, ICD-9-CM 820-821 and 823, and operation

codes 81.51-81.54), cancer (ICD-9-CM 140-208), diabetes

(ICD-9-CM 250), hyperlipidemia (ICD-9-CM

270), heart failure 9-CM 428), stroke

(ICD-9-CM 430-438), atrial fibrillation (ICD-(ICD-9-CM 427.31)

and bedsores (ICD-9-CM 707). All comorbidities were

defined before the date of PE diagnosis.

We also calculated the duration and dosage of morphine

use for each patient. Based on a previous study

(11), we defined morphine use as ‘current’ if the prescription

duration covered the index date or ended less

than, or equal to, 30 days before the index date.Moreover,

morphine use was defined as ‘recent’ if the prescription

ended 31–180 days before the index date, and

morphine use was labeled ‘past’ if the prescription

ended more than 180 days prior.

All statistical analyses were performed using SAS 9.2

software for windows (SAS Institute, Cary, NC, USA),

and the significance level was set at 0.05. The chisquared

test was used to assess the categorical variables

between the case and control groups. The odds ratio

and 95% confidence intervals (95% CIs) for PE were

assessed using logistic regression.We also assessed the

association between PE and morphine dosage, stratified

by the median.

Results

Data of 1555 DVT patients were used in this study: 174

patients were assigned to the case group, and 1381

patients were assigned to the control group. PE

age: 64.9 and 61.6 years, respectively; standard deviation = 16.5 and 16.7, respectively; P = 0.02, data not

shown). Female patients outnumbered their male PE counterparts. Compared with the control patients, PE patients had a history of more comorbidities, but the difference was not statistically significant (Table 1). Compared with the study patients who did not

receive morphine, patients who used morphine within 30 days of the index date had a 4.54-fold higher risk for developing PE (95% CI = 2.30–8.97, Table 2). The likelihood of PE development in relation to morphine

dosage levels is shown in Table 3. Overall, there was no significant difference among the categories stratified by cumulative dosage, average dosage and the duration of morphine use. However, patients who used morphine within 30 days before the index date had a significantly higher risk: the risks for PE increased from 3.82 to 5.77 and from 3.73 to 5.91 with the cumulative dosage (from 1–199 to ≥200 mg) and average dosage (from 0.01–8.99 to ≥9.00 mg/d), respectively.

Discussion

Our results indicated that current morphine use is associated with the incidence of PE in DVT patients. The risk of PE increased with augmented morphine dosage only in patients treated with morphine within the past 30 days.

Using the NHIRD of Taiwan, our group has previously demonstrated that both rheumatoid arthritis

and systemic lupus erythematosus increase the risk of DVT and PE (12, 13). The risk of DVT and PE also increases significantly in spinal cord injury patients, especially within 3 months after the occurrence of

spinal cord injury (14). Frequent asthma exacerbation and hospitalization are significantly associated with PE

(15). Moreover, patients with idiopathic VTE have a 6.89% incidence of subsequent cancer diagnosis in Taiwan (16).Here, our analysis with the same database further shows that PE is associated with current morphine use in DVT patients.

The risk assessment and appropriate treatment selection for PE patients remain a challenge (17). Clinical probability assessments help identify patients with low probability. For them, a PE diagnosis can be excluded solely by a negative result in a plasma D-dimer test (18, 19). The diagnosis is usually confirmed

with a chest CT showing PE. The foundation of treatment is anticoagulation. Therefore, PE patients should be assessed for considering the addition of an advanced treatment, such as thrombolysis or possibly an embolectomy (19–21). A study conducted in

Taiwan suggested considering the initial appearance of tachycardia or the presence of a chronic pulmonary disease elevating the white blood cell count and an increased D-dimer concentration upon admission as ways to identify the risk of a fatal outcome in patients with acute PE (22). Our findings that current morphine use is a risk factor for PE should be considered as an addition to the clinical checklist to facilitate the early diagnosis of PE in DVT patients.

Although the risk of PE development diminishes 30 days after morphine treatment has ceased, DVT patients currently under morphine administration should be monitored to prevent the short-term risk of PE development during morphine use. A recent study reported that patients receiving extended-release epidural morphine (EREM) after a total hip arthroplasty

had a slightly higher incidence of PE (23). In contrast, the use of EREM following a lower extremity joint arthroplasty was demonstrated to be associated with a significant reduction in the incidence of PE (24).

The discrepancy between these two studies might be resolved if the length of time where morphine was administered is introduced as a factor. In a comparative proteomic study on acute PE using a rat as the test subject, the Na+/H+ exchanger was identified to display

differential expression patterns after PE (25). A further investigation is required to enhance our understanding of the relationship between morphine and the reversal of blood clot formation at the molecular level.

The strengths of our study include the use of

population-based data and the evaluation of NHIRD records, rather than data obtained from self-reported drug use. The high accuracy and validity of diagnosis

in PE (12–16), cardiology-related and autoimmune

disease in NHIRD have been demonstrated (26), indicating the validity and accuracy about the ICD-9 codes

in NHIRD of Taiwan. However, the limitations of

this study are as follows: First, the NHIRD lacks important data, such as detailed demographic information

on smoking habits, alcohol consumption, body mass index, socioeconomic status and a family history of systemic diseases. These are all potential risk factors for PE development, and each factor is indirectly associated with morphine use. Therefore, we were unable to

correlate an increased morphine dosage with inactivity and/or malnutrition because the NHIRD does not contain lifestyle data.

However, because the NHIRD covers a highly representative sample of Taiwan’s general population, and

because the insurance reimbursement policy is universal, these factors are unlikely to have affected morphine prescription allotment in the sample group. Second, because DVT might influence mobility and muscle strength, we used the nested case-control method to eliminate confounding factors caused by DVT. Yet evidence derived from a nested case-control study is generally of lower quality than that obtained from randomized trials. This is because a nested case-control study design is subject to several biases stemming from adjustments made for confounding variables. Despite a

meticulous study design that adequately controls confounding factors, a potential key limitation of this study

is that a bias could remain if unknown confounders are present. Third, the diagnoses recorded in the NHI

claims are primarily used for administrative billing purposes and have not been verified for scientific purposes.

We were unable to contact patients directly to inquire about their morphine use because all beneficiaries listed in theNHIRDare protected by anonymity.Wewere also unable to consider morphine prescriptions issued

of cumulative dosages, and may have subsequently weakened the observed association. However,

the data we obtained onmorphine prescription and PE diagnosis were reliable.

Our results indicate that the PE incidence is associated with current morphine treatment. However, the

effects of morphine appear to be reversible because this trend was not observed in patients who ceased morphine use more than 30 days before the index date of

the study. However, further large population-based studies, or large-scale randomized clinical trials, are required to confirmthese findings before any definitive conclusions can be drawn.