Background Few studies have investigated the relationship between carbon monoxide (CO) poisoning and risk of deep vein thrombosis (DVT) and pulmonary embolism (PE). Therefore, we conducted a nationwide longitudinal cohort study in Taiwan to determine whether patients with CO poisoning are associated with increased risk of developing DVT and PE.
Methods This study investigated the incidence and risk factors for DVT and PE in 8316 patients newly diagnosed with CO poisoning from the Taiwan National Health Insurance Research Database between 2000 and 2011. The comparison cohort contained 33 264 controls without CO poisoning from the general population. Follow-up was initiated on the date of initial diagnosis of CO poisoning and continued until the date of a DVT or PE event, censoring or December 31, 2011. Cox proportional hazard regression models were used to analyse the risk of DVT and PE according to sex, age and comorbidities.
Results The incidences of DVT and PE were higher in the patients with CO poisoning than in the controls (5.67 vs 1.47/10 000 person-years and 1.97 vs 1.02/10 000 person-years, respectively). After adjusting for age, sex and comorbidities, the patients with CO poisoning were associated with a 3.85-fold higher risk of DVT compared with the comparison cohort, and non-significantly associated with risk of PE. CO poisoning patients with a coexisting comorbidity or acute respiratory failure were associated with significantly and substantially increased risk of DVT.
Conclusion Risk of DVT is significantly higher in patients with CO poisoning than in the general population.
- Clinical epidemiology
- Cohort studies
- VASCULAR DISEASE
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Carbon monoxide (CO) poisoning is a major toxicological cause of morbidity and mortality worldwide. CO concentrations ≥667 ppm can cause ≤50% of the body's haemoglobin to convert to carboxyhaemoglobin, which interferes with the transport of oxygen to tissues by haemoglobin.1 CO binds to haemoglobin with an affinity 200–240 times that of oxygen. CO poisoning related to tissue hypoxia has been well established.2 CO also causes tissue damage through oxidative stress, free radical production, inflammation and apoptosis.3 ,4
Deep vein thrombosis (DVT) is the formation of blood clots in the deep vein, with thrombi predominantly occurring in the legs. It is a serious condition that can lead to blockage of the major pulmonary artery when the blood clots travel to the lung. Pulmonary embolism (PE) is a potentially life-threatening disorder. Together, DVT and PE constitute venous thromboembolism (VTE). VTE is associated with Virchow's triad: hypercoagulability, stasis and endothelial dysfunction.5 VTE is not a benign disease, and has a 30-day cases death rate of 11–30%.6–9
The symptoms of CO poisoning are non-specific.10 Mild CO poisoning tends to cause headache, dizziness and neuropsychological impairment.11 Severe CO poisoning can cause confusion, unconsciousness and even death.12 Recent studies on CO intoxication have focused on outcomes of neurological injury and cardiac damage.13–18 Other studies have shown that CO poisoning causes inflammation and hypoxia, which might induce endothelial dysfunction and hypercoagulability.19–21 Inflammation might also play a role in VTE.22 However, few studies have evaluated the relationship between CO poisoning and DVT or PE. Therefore, we conducted a nationwide retrospective study to investigate the risk of DVT and PE in patients with CO poisoning.
The Taiwan National Health Insurance (NHI) programme was initiated in March 1995 and covers >99% of Taiwan's population (23.74 million people).23 The NHI Research Database (NHIRD) is derived from the payment system of the NHI Administration (NHIA), managed by the National Health Research Institute (NHRI). Confidentiality is maintained according to the directives of the NHIA in Taiwan. All of the identifiers of individual patients are deleted by the NHIA before data are transferred to the NHIRD. The NHIRD is used for administrative and research purposes. Institutional review board approval is preapproved by the NHRI for de-identified data. Details on the generation, monitoring and maintenance of the NHIRD are published online by the NHRI. Diagnostic codes in the NHIRD are based on the International Classification of Diseases, ninth revision, Clinical Modification (ICD-9-CM). Previous studies have described the high accuracy and validity of ICD-9-CM diagnoses in the NHIRD.24 ,25 This study was approved from full review by the institutional research ethics committee (CMU-REC-101-012).
Adults with newly diagnosed CO poisoning (ICD-9-CM Code 986) were identified from the inpatient claims during 2000–2010 as the CO poisoning cohort. The date of the first admission for CO poisoning was used as the index date. Patients diagnosed with DVT (ICD-9-CM Code 453.8) or PE (ICD-9-CM Code 415.1, excluding ICD-9-CM Code 415.11) at the baseline were excluded. Participants aged <20 years and those without information on age and sex were also excluded. For every CO poisoning patient, four controls were randomly selected from all NHI beneficiaries without CO poisoning, and the exclusion criteria used for selecting the CO poisoning cohort were applied. The controls were frequency-matched with the CO poisoning patients according to age (every 5 years), sex and year of hospitalisation. Overall, 8316 CO poisoning patients and 33 264 non-CO poisoning controls were followed up until a diagnosis of DVT or PE, loss to follow-up, death, withdrawal from the NHI programme, or the end of 2011.
The baseline comorbidity history for each participant was determined from the inpatient claims data. Several well-known risk factors for DVT or PE, including atrial fibrillation (ICD-9-CM Code 427.31), hypertension (ICD-9-CM Codes 401–405), diabetes (ICD-9-CM Code 250), cerebral vascular disease (CVA; ICD-9-CM Codes 430–438), heart failure (ICD-9-CM Code 428), all cancers (ICD-9-CM Codes 140–208), pregnancy (ICD-9-CM Codes 640.x1–676.x1, 640.x2–676.x2, 650–659 and procedure codes 72–74) and lower leg fracture or surgery (ICD-9-CM Codes 820–823 and procedure codes 81.51, 81.52, 81.53, and 81.54), were considered. Acute respiratory failure (ARF; ICD-9-CM Code 518.81) was also considered and identified according to its diagnoses in the hospitalisation records within 3 days of the patient's index date. We also evaluated the risks of DVT and PE for the CO poisoning patients who received hyperbaric oxygen therapy (Procedure Code 93.95).
A χ2 test was used to determine the differences in categorical demographic variables and comorbidities between the CO poisoning and control cohorts. A Student t test was used to compare the mean ages of the two cohorts. The cumulative incidence curves of DVT or PE for the two cohorts were compared using the Kaplan-Meier method and log-rank test. The overall and sex-speciﬁc, age-speciﬁc and comorbidity-speciﬁc incidence rates (per 10 000 person-years) were calculated for each cohort. The relative HRs and 95% CIs for DVT or PE, according to each variable, were calculated using Cox proportional hazard regression analysis. A multivariable Cox proportional hazard regression analysis was performed to estimate the adjusted HRs and 95% CIs for DVT or PE development in the CO poisoning cohort compared with the control cohort, adjusting for age, sex and comorbidities of hypertension, diabetes, CVA, heart failure, all cancers, pregnancy and lower leg fracture or surgery. All analyses were conducted using SAS statistical software (V.9.2 for Windows; SAS Institute, Inc, Cary, North Carolina, USA). Two-tailed p<0.05 was considered significant.
Demographic characteristics and comorbidities of the CO poisoning and control patients
Table 1 lists the characteristics of the patients in the CO poisoning and control cohorts. The majority of the patients were aged <40 years (58%) and more than half were men (52.3%). The mean age was 39.6±13.8 years in the CO poisoning cohort and 39.5±14.1 years in the control cohort. Comorbidities at the baseline were more prevalent in the CO poisoning cohort than in the control cohort (p<0.01).
Incidences and adjusted HRs of DVT and PE according to sex, age and comorbidities in the CO poisoning patients compared with the control patients
The overall incidence of DVT was 3.82-fold higher in the CO poisoning cohort than in the comparison cohort (5.67 and 1.47/10 000 person-years, respectively; table 2). After adjusting for covariates, the risk of developing DVT was significant in the patients with CO poisoning (adjusted HR=3.85; 95% CI 2.17 to 6.83). In both cohorts, the incidence rate of DVT was marginally higher in women than in men. The adjusted HR for DVT was higher in women (adjusted HR=3.75, 95% CI 1.71 to 8.22) and men (adjusted HR=3.86, 95% CI 1.65 to 8.99) in the CO poisoning cohort compared with in the comparison cohort. The incidence of DVT increased with age in both cohorts. However, the age-specific relative risk of DVT in the CO poisoning patients compared with in the control patients was highest in the patients aged <40 years (adjusted HR=13.8; 95% CI 3.80 to 49.9). In both cohorts, the incidence of DVT increased in the presence of a comorbidity. We analysed the association between CO poisoning and the risk of DVT stratifying by comorbidity and found a 5.39-fold risk of DVT was significantly observed in patients without comorbidity (adjusted HR=5.39, 95% CI 2.52 to 11.5). The overall incidence of PE in the patients with CO poisoning and control patients exhibited non-significant differences (1.97 vs 1.02/10 000 person-years).
Joint effects of comorbidity or respiratory failure or hyperbaric oxygen therapy on risk of CO poisoning-associated DVT and PE
Table 3 lists the joint effects of CO poisoning and comorbidity, respiratory failure or hyperbaric oxygen therapy on the risk of DVT and PE. We observed significantly higher risk of DVT in patients with CO poisoning and associated comorbidity (adjusted HR=6.97, 95% CI 3.05 to 16.0), and in patients with CO poisoning and respiratory failure (adjusted HR=5.17, 95% CI 1.55 to 17.2), compared with patients without CO poisoning and comorbidity or respiratory failure. Compared to those without CO poisoning and hyperbaric oxygen therapy, patients with CO poisoning and hyperbaric oxygen therapy were 6.89-fold more likely to develop DVT (95% CI 1.70 to 6.07), followed by patients with CO poisoning and without hyperbaric oxygen therapy (HR=3.21, 95% CI 1.70 to 6.07) for developing DVT.
The patients with CO poisoning and comorbidity were associated with a substantially higher risk of PE than patients without CO poisoning and comorbidity (adjusted HR=5.07, 95% CI 1.76 to 14.6).
Cumulative incidences of DVT and PE in patients with CO poisoning and control patients
The Kaplan-Meier graph in figure 1A illustrates that the cumulative incidence of DVT was higher in the CO poisoning cohort than in the comparison cohort (log-rank test p<0.001). The cumulative incidence of PE in the CO poisoning and comparison cohorts exhibited non-significant differences (figure 1B, log-rank test p=0.11).
According to our research, this study is the first nationwide population-based cohort study to demonstrate that patients with CO poisoning are associated with higher risk of developing DVT (3.85-fold) than control patients from the general population. The patients with CO poisoning were associated with a significantly higher proportion of comorbid diseases compared with the control patients. However, CO poisoning remained an independent risk factor for developing DVT after adjusting for sex, age and comorbidities. The risk of PE in the patients with CO poisoning and control patients exhibited non-significant differences.
The epidemiological causes of increased risk of DVT in patients with CO poisoning are unclear. Exogenous exposure to high amounts of CO can result in toxic effects, causing hypoxia and inflammation.10 ,26 CO enters the blood to induce hypoxia through the formation of carboxyhaemoglobin, which causes a leftward shift in the oxyhaemoglobin dissociation curve, and binds to protein to impair mitochondrial function.12 CO also causes inflammation by increasing the levels of cytosolic and oxygenase-1 protein, and inducing the release of myeloperoxidase, protease and reactive oxygen species, which results in oxidative stress.4 The pathogenesis of VTE is reportedly associated with inflammation and oxidative stress.27
The patients with CO poisoning who presented with ARF were associated with substantially increased risk of DVT compared with the control patients. This finding supports previous studies’ observations of increased risk of DVT in patients with chronic obstructive pulmonary disease and ARF.28 ARF-induced hypoxaemia, associated with hypercoagulability, endothelial dysfunction29 and immobilisation, might explain our observations of markedly increased risk of DVT in CO poisoning patients with ARF. The patients with CO poisoning receiving hyperbaric oxygen therapy exhibited a considerable risk of DVT, which may be associated with disease severity. The CO poisoning patients with coexisting comorbidities were also associated with substantially increased risk of DVT compared with the comparison patients. Therefore, clinicians should consider providing DVT prophylactic medication to CO poisoning patients with ARF or coexisting comorbidities.
The incidence of DVT was marginally higher in women than in men in the CO poisoning and comparison cohorts. Previous studies have provided inconsistent results on sex differences in the incidence of DVT.30 ,31 The incidence of DVT increased with age in CO poisoning and comparison cohorts. However, in the patients aged <40 years risk of DVT was significantly higher in the CO poisoning patients than in the comparison cohort. Clinicians should thus be aware of increased risk of DVT and provide multidisciplinary care to adults aged <40 years with CO poisoning.
The strengths of our study are its nationwide population-based design and longitudinal evaluation of the risk of DVT and PE in CO poisoning patients from an Asian population. Because each resident in Taiwan is assigned a unique personal identification number, all study participants could be traced through the records of the NHI for the entire follow-up. Thus, our findings can be generalised to the entire population of Taiwan.
However, this study also has several limitations. First, the NHIRD does not provide detailed information on potential confounding factors, such as smoking, body mass index and physical activity. The absence of drug data, such as those on hormone replacement therapy, the use of contraceptive drugs and anticoagulant treatments, might also have influenced the primary study outcomes. Lack of laboratory data (such as HbCO) and imaging information (such as perfusion lung scan) became another limitation, of which we could not evaluate the disease severity for CO poisoning. Healthcare claim data can potentially be associated with misclassification bias for primary outcomes. However, the auditing mechanisms applied by the NHIA minimise diagnostic uncertainty and misclassification.32 ,33 Finally, despite our meticulous study design, in which the confounding factors were adequately controlled, data derived from a retrospective cohort study are generally of lower statistical quality than those derived from randomised trials because of potential biases.
In conclusion, the results from our nationwide population-based cohort study, examining 8316 CO poisoning patients with a follow-up duration of approximately 41 000 person-years, show that patients with CO poisoning are associated with 3.85-fold higher risk of DVT compared with the general population, but are non-significantly associated with risk of PE. These findings emphasise the requirement for a multidisciplinary approach for the management of potential risk factors for DVT in patients with CO poisoning. However, additional studies are required to elucidate the biological mechanisms underlying increased risk of DVT development in patients with CO poisoning.
What is already known on this subject
Few studies have investigated the relationship between carbon monoxide poisoning and risk of deep vein thrombosis and pulmonary embolism.
What this study adds
The incidences of deep vein thrombosis and pulmonary embolism were higher in the patients with carbon monoxide poisoning than in the controls (5.67 vs 1.47/10 000 person-years and 1.97 vs 1.02/10 000 person-years, respectively).
Contributors W-S C and C-H K were involved in the conception/design; C-H K was involved in the collection and/or assembly of data. All authors were involved in the data analysis and interpretation and in the manuscript writing and gave final approval of the manuscript. C-H K is the guarantor, took responsibility for the integrity of the work as a whole, from inception to the published article.
Funding This work was supported by grants from China Medical University (CMU102-BC-2); Taiwan Ministry of Health and Welfare Clinical Trial and Research Center of Excellence (MOHW103-TDU-B-212-113002), Health, and welfare surcharge of tobacco products, China Medical University Hospital Cancer Research Center of Excellence (MOHW104-TD-B-111-03, Taiwan).
Competing interests None.
Ethics approval This study was approved by full review by the Institutional Research Ethics Committee (CMU-REC-101-012).
Provenance and peer review Not commissioned; externally peer reviewed.