SEARCH

SEARCH BY CITATION

Keywords:

  • coffee consumption;
  • risk factor;
  • venous thromboembolism

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure of Conflict of Interests
  7. References

Summary. Background: Several studies have investigated the association between coffee consumption and cardiovascular disease, but little is known about coffee intake and the risk of venous thromboembolism (VTE). Objective: The aim of this prospective cohort study was to investigate the association between coffee consumption and the risk of incident VTE in a general population. Methods: Information about coffee consumption habits was obtained with a self-administered questionnaire in 26 755 subjects, aged 25–97 years, who participated in the fourth survey of the Tromsø study (1994–1995). Incident VTE events were registered until the end of follow-up, 1 September 2007. Results: There were 462 incident VTE events (1.60 per 1000 person-years, 95% confidence interval [CI] 1.46–1.75) during a median of 12.5 years of follow-up. A daily consumption of three to four cups was borderline associated (hazard ratio [HR] 0.70; 95% CI 0.48–1.02) and a daily consumption of five to six cups (HR 0.67; 95% CI 0.45–0.97) was significantly associated with reduced risk of VTE as compared with coffee abstainers in multivariable analysis adjusted for age, sex, body mass index (BMI), smoking status, physical activity, diabetes, history of cardiovascular disease and cancer. Similar risk estimates were found for provoked and unprovoked VTE, and in sex-stratified analyses. Conclusion: Our findings suggest a possible U-shaped relationship between coffee consumption and VTE, and that moderate coffee consumption may be associated with a reduced risk of VTE. However, more studies are needed to establish whether moderate coffee consumption is inversely associated with the risk of VTE.

Venous thromboembolism (VTE), consisting of deep vein thrombosis (DVT) and pulmonary embolism (PE), is a disease with serious short-term and long-term complications, such as post-thrombotic syndrome, pulmonary hypertension, and potentially fatal outcome [1,2]. The annual incidence of VTE is 1–2 per 1000 person-years in developed countries [1,3,4], and it is the leading cause of preventable in-hospital death in the USA [2]. VTE is a multicausal disease, the incidence of which increases markedly with age [1,4,5]. Despite numerous known risk factors, 25–50% of the events still occur in the absence of predisposing factors [2,6]. In contrast to that of arterial cardiovascular disease (CVD) (e.g. myocardial infarction [MI]), the incidence of VTE has not declined during recent decades [1]. The relationship between lifestyle factors and arterial CVD has been extensively examined, whereas limited knowledge exists on the association between lifestyle factors and risk of VTE.

Coffee is one of the most widely consumed beverages in the world, and has been associated with both health benefits and health risks [7]. Epidemiologic studies have reported diverging results regarding the association between coffee consumption and risk of CVD [8–11], but more recent cohorts have shown an inverse association [8–11]. Recently, a J-shaped curve was suggested for the association between coffee intake and acute coronary syndrome (ACS) [12], where moderate coffee consumption was associated with reduced risk of ACS.

To our knowledge, only one prospective cohort study, including older women only, has investigated the association between coffee consumption and venous thrombosis [13]. This study reported a weak inverse association between coffee consumption and VTE, but the association disappeared after adjustments for body mass index (BMI) and diabetes [13]. Thus, our current knowledge of the impact of coffee consumption on the risk of VTE is limited. The aim of the present study was to explore the association between coffee consumption and the risk of future VTE, and to test whether this relationship was linear, in a large prospective cohort recruited from a general population.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure of Conflict of Interests
  7. References

Study population

Subjects were recruited from the fourth Tromsø study, a single-center, prospective, population-based health study that was carried out in 1994–1995. All inhabitants of the municipality of Tromsø, aged 25 years or older, were invited to participate, and 77% of the eligible population (n = 27 158) participated. The study was approved by the regional committee of research ethics, and all subjects gave their written consent to participate. Participants were excluded for the following reasons: 300 did not give their consent to medical research, 43 were not officially registered as inhabitants of the municipality of Tromsø at the date of enrollment, 47 had a known history of VTE, and 13 had missing values for coffee consumption. Thus, a total of 26 755 participants were included in the study, and were followed from the date of enrollment in 1994–1995 until the end of follow-up, 1 September 2007.

Baseline measurements

Baseline information was collected by physical examination, blood sampling and self-administered questionnaires at a single time-point (date of inclusion). Height and weight were measured with subjects wearing light clothes and no shoes. BMI was calculated as weight in kilograms divided by the square of height in meters (kg m−2). Non-fasting blood samples were collected from an antecubital vein, and serum was prepared by centrifugation after 1 h rest at room temperature, and further analyzed at the Department of Clinical Chemistry, University Hospital of North Norway. Serum total cholesterol was analyzed by the CHOD-PAP method (Boeringer Mannheim, Mannheim, Germany). Baseline information on coffee consumption, diabetes, smoking status, prior CVD (MI, angina pectoris and stroke), daily caloric intake, physical activity and current hormone therapy was collected with a self-administered questionnaire. The coffee consumption questions were: ‘How many cups of boiled coffee (coarsely ground coffee for brewing) do you usually drink daily?’ and ‘How many cups of coffee, other than boiled, do you usually drink daily?’ Physical activity was defined as exercise with sweat production and breathlessness for ≥ 1 h per week during leisure time. Current hormone therapy was defined as current use of oral contraceptives or current use of estrogen supplementation (tablets or patches). Information on cancer was obtained from the Norwegian Cancer Registry, and cancer exposure was defined as all diagnoses of cancer within the last 10 years prior to the baseline inclusion date. Data on daily caloric intake were available for a subpopulation of participants aged < 70 years (n = 17 141), and were calculated from information on dietary habits given in the questionnaire [14].

Identification and validation of VTE

All incident VTE events during follow-up (from date of examination in 1994–1995 until 1 September 2007) were identified by searching the hospital discharge diagnosis registry, the autopsy registry and the radiology procedure registry at the University Hospital of North Norway, as previously described [5]. The University Hospital of North Norway is the only hospital in the region, and all hospital care and diagnostic radiology is provided exclusively by this hospital.

The medical records for each potential VTE case were reviewed by trained personnel, who were blinded to all the baseline variables, including coffee consumption habits. Unclear cases or disputes were resolved by discussion with a senior consultant expert on VTE diagnosis (J.B.H.). For subjects derived from the hospital discharge diagnosis registry and the radiology procedure registry, an episode of VTE was verified and recorded as a validated outcome when all four of the following criteria were fulfilled: (i) objectively confirmed by diagnostic procedures (compression ultrasonography, venography, spiral computed tomography, perfusion–ventilation scan (moderate and high probability of PE required), pulmonary angiography or autopsy); (ii) the medical record indicated that a physician had made a diagnosis of DVT or PE; (iii) signs and symptoms consistent with DVT or PE were present; and (iv) the patient had undergone therapy with anticoagulants (heparin, warfarin or a similar agent), thrombolytic therapy or surgery unless contraindications to anticoagulant treatment were specified in the medical record. Isolated calf vein DVT was diagnosed by compression ultrasonography alone or accompanied by venography if the diagnosis was insecure and/or there were inadequate technical conditions (e.g. obesity). For subjects derived from the autopsy registry, a VTE event was recorded as an outcome when the autopsy registry indicated VTE as cause of death or as a significant condition contributing to death.

A VTE event was classified as provoked (≥ 1 provoking factors) or unprovoked (no provoking factors) by the presence of provoking factors at the time of diagnosis. Provoking factors were: recent surgery or trauma within the previous 8 weeks, acute medical conditions (acute MI, ischemic stroke or major infectious disease), active cancer, marked immobilization (bed rest for > 3 days, wheelchair use, or long-distance travel exceeding 4 h within the last 14 days prior to the event) or other provoking factors specifically described by a physician in the medical record (e.g. intravascular catheter).

Statistical analyses

For each participant, person-years were accrued from the date of enrollment in the Tromsø study (1994–1995) to the date of the first VTE event, the date when the participant died or moved from the municipality of Tromsø, or the end of follow-up, 1 September 2007. Subjects who moved from the municipality of Tromsø (n = 3716) or died (n = 3000) during the study period were censored.

Statistical analyses were performed with stata version 11.0 (Stata Corporation, College Station, TX, USA). The distribution of baseline characteristics was adjusted for age (with crude standard deviations), and tests for linear trends were carried out with logistic and linear regression. Age-adjusted and sex-adjusted incidence rates (IRs) with 95% confidence intervals (CIs) were calculated by direct standardization as number of events per 1000 person-years, with the age distribution of the whole cohort as the standard population. The association between daily coffee consumption and the risk of total, provoked and unprovoked VTE was assessed with Cox proportional hazard regression models to estimate age- and sex-adjusted and multivariable-adjusted hazard ratios (HRs) for VTE with 95% CIs. The two questionnaire variables on coffee consumption (by type of coffee) were combined into a daily coffee consumption variable, which was used as our main exposure variable divided into five categories: no cups per day, one to two cups per day, three to four cups per day, five to six cups per day and more than six cups per day. The coffee abstainers (no cups per day) served as the reference group. To correct for potential confounding, other risk factors and concomitant diseases possibly related to both coffee consumption and venous thrombosis were included in a multivariable model. Hence, in the multivariable analyses, the HRs were adjusted for age, sex, BMI, smoking status (current smoker yes/no), physical activity, prior history of CVD, cancer and self-reported diabetes. Adjustments for daily caloric intake in addition to the above-mentioned factors were performed in a model including a subgroup of participants aged < 70 years only. Further adjustments for estrogen use (as a three-level variable: men, women not on estrogens, and women on estrogens) were also conducted. Statistical interactions of coffee consumption with age or sex were tested by including cross-product terms in the proportional hazard models. There were no statistical interactions between coffee consumption and age or sex. The proportional hazard assumption was verified by evaluating the parallelism between the curves of the log–log survival function for categories of total coffee consumption. In addition, a test of the proportional hazard assumption using Schoenfeld residuals was performed for all of the relevant variables.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure of Conflict of Interests
  7. References

Table 1 shows the characteristics of VTE patients at the time of the event. In total, there were 462 validated incident VTE events during 289 338 person-years of follow-up (median follow-up time 12.5 years). The overall crude IR was 1.60 per 1000 person-years (95% CI 1.46–1.75). Of the events, 64.3% (n = 297) were DVT, and the remaining 35.7% (n = 165) were PE with or without concurrent DVT (Table 1). Moreover, 41.8% (n = 193) of the events were classified as unprovoked. Active cancer was the most common provoking factor for VTE. The proportions of DVT and PE, provoked and unprovoked VTE and the presence of clinical risk factors and provoking factors did not differ between sexes (data not shown).

Table 1.   Characteristics of the incident venous thromboembolism (VTE) events (n = 462) during follow-up. The Tromsø study 1994/1995–2007
 % (n)
  1. *Only women were included in the analysis. †VTE in first-degree relative before the age of 60 years. ‡Includes other diseases within the previous year (myocardial infarction, ischemic stroke, heart failure, inflammatory bowel disease, chronic infections, chronic obstructive pulmonary disease or myeloproliferative disorders). §Immobility includes bed rest for > 3 days, long-duration travel by car, boat, train or air of > 4 h within the last 14 days, or other type of immobilization. ¶Other provoking factor described by a physician in the medical record (e.g. intravascular catheter).

Men47.4 (219)
Deep vein thrombosis64.3 (297)
Pulmonary embolism35.7 (165)
Unprovoked41.8 (193)
Clinical risk factors
 Estrogen*14.4 (35)
 Pregnancy/puerperium*1.2 (3)
 Heredity†2.8 (13)
 Other medical conditions‡21.6 (100)
Provoking factors
 Surgery17.1 (79)
 Trauma6.7 (31)
 Acute medical conditions15.1 (70)
 Cancer22.9 (106)
 Immobility§19.0 (88)
 Other¶4.1 (19)

The distribution of age-adjusted characteristics of subjects across categories of daily coffee consumption at baseline is shown in Table 2. Age, BMI and total cholesterol levels increased significantly across categories of higher coffee consumption. The proportion of men and smokers increased, while the degree of physical activity and current use of hormone therapy (women only) decreased with increasing coffee consumption.

Table 2.   Baseline characteristics across categories of coffee consumption*. The Tromsø Study 1994/1995–2007
Characteristics0 cups per day1–2 cups per day3–4 cups per day5–6 cups per day> 6 cups per dayP for trend
  1. BMI, body mass index; CVD, cardiovascular disease. *Values are given as age-adjusted means ± crude standard deviations or as age-adjusted percentages with absolute numbers in parentheses. †Includes history of myocardial infarction, angina pectoris or stroke. ‡The percentages are based on the female population. §Based on a subpopulation of participants aged < 70 years (n = 17 141).

Number of subjects26712873658171787452
Number of events3651126118131
Men39.6 (1059)39.4 (1133)41.7 (2742)47.0 (3372)59.2 (4411)< 0.001
Age (years)38.5 ± 14.346.9 ± 17.149.7 ± 16.248.3 ± 14.546.0 ± 12.6< 0.001
BMI (kg m−2)25.2 ± 4.124.9 ± 3.825.0 ± 3.825.2 ± 3.825.4 ± 3.8< 0.001
Total cholesterol (mm)5.75 ± 1.205.85 ± 1.305.99 ± 1.356.09 ± 1.296.25 ± 1.27< 0.001
Self-reported diabetes1.1 (30)1.3 (69)1.0 (137)1.2 (140)1.0 (103)0.50
Self-reported prior CVD†2.8 (103)2.8 (223)2.6 (535)2.9 (509)3.1 (404)0.06
Current smoking15.6 (494)16.0 (476)24.2 (1560)39.3 (2790)58.0 (4341)< 0.001
Hormone therapy‡23.7 (271)23.9 (265)20.8 (480)18.3 (449)14.9 (299)< 0.001
Physical activity30.0 (1005)31.1 (937)30.5 (1955)28.9 (2077)26.4 (2083)< 0.001
Cancer1.3 (37)1.4 (74)1.2 (160)1.3 (163)1.3 (129)0.91
Daily caloric intake§1775 ± 5131803 ± 4961843 ± 5121890 ± 5161997 ± 553< 0.001

IRs and HRs for total, provoked and unprovoked VTE across categories of increasing coffee consumption are presented in Table 3. The age- and sex-adjusted IR was highest among the coffee abstainers (IR 2.14 per 1000 person-years; 95% CI 1.54–2.96). Moderate coffee consumption was inversely associated with risk of VTE (Table 3). Drinking three to four cups per day (HR 0.70, 95% CI 0.48–1.02) was borderline associated and drinking five to six cups per day (HR 0.67, 95% CI 0.45–0.97) was significantly associated with a reduced risk of VTE as compared with coffee abstainers in analyses adjusted for age, sex, BMI, smoking status, physical activity, cancer, prior CVD and diabetes. The inverse association of coffee consumption was attenuated for subjects drinking more than six cups of coffee daily (multivariable HR 0.85; 95% CI 0.58–1.24). Similar risk estimates were found for provoked and unprovoked VTE in separate analyses (Table 3). Inclusion of daily caloric intake in addition to the other adjustments among participants aged < 70 years did not significantly change the risk estimates (Table 3). Further adjustments for current hormone therapy did not significantly alter the risk of VTE (data not shown). The inverse association between moderate coffee intake and risk of VTE was similar in men and women in sex-stratified analysis (data not shown), and for boiled and non-boiled coffee intake (data not shown).

Table 3.   Incidence rates (IRs) and hazard ratios (HRs) for venous thromboembolism (VTE) by categories of daily coffee consumption. The Tromsø study 1994/95–2007
 Person-yearsEventsIR*†HR†Multivariable HR‡Multivariable HR§
  1. *Incidence rate per 1000 person-years. †Adjusted for age and sex, with 95% confidence intervals. ‡Adjusted for age, sex, body mass index (BMI), smoking status, physical activity, diabetes, history of cardiovascular disease (CVD) and cancer, with 95% confidence intervals. §Adjusted for age, sex, BMI, smoking status, physical activity, diabetes, history of CVD, cancer and daily caloric intake, with 95% confidence intervals. Includes only participants aged < 70 years. ¶Includes a subpopulation of participants aged < 70 years (n = 17 141).

Total VTE
 0 cups per day27 298362.14 (1.54–2.96)Ref.Ref.
 1–2 cups per day29 808511.59 (1.21–2.09)0.77 (0.50–1.18)0.78 (0.51–1.21)
 3–4 cups per day70 4351261.51 (1.27–1.80)0.70 (0.48–1.02)0.70 (0.48–1.02)
 5–6 cups per day79 0141181.40 (1.17–1.68)0.68 (0.47–0.99)0.67 (0.45–0.97)
 > 6 cups per day82 7831311.74 (1.46–2.06)0.88 (0.61–0.99)0.85 (0.58–1.24)
 P for trend0.960.71
Provoked VTE
 0 cups per day27 199211.21 (0.79–1.85)Ref.Ref.
 1–2 cups per day29 646250.78 (0.53–1.15)0.62 (0.35–1.12)0.61 (0.34–1.10)
 3–4 cups per day70 102760.88 (0.70–1.10)0.70 (0.43–1.14)0.67 (0.41–1.09)
 5–6 cups per day78 635680.81 (0.64–1.03)0.66 (0.40–1.08)0.62 (0.38–1.01)
 > 6 cups per day82 401791.05 (0.84–1.31)0.92 (0.57–1.49)0.84 (0.51–1.38)
 P for trend 0.500.84
Unprovoked VTE
 0 cups per day27 190150.95 (0.57–1.57)Ref.Ref.
 1–2 cups per day29 622260.86 (0.58–1.26)0.98 (0.52–1.86)1.06 (0.55–2.04)
 3–4 cups per day69 936500.65 (0.49–0.85)0.70 (0.39–1.25)0.75 (0.41–1.36)
 5–6 cups per day78 539500.61 (0.46–0.80)0.71 (0.40–1.26)0.74 (0.41–1.35)
 > 6 cups per day82 193520.70 (0.53–0.92)0.84 (0.47–1.49)0.85 (0.46–1.56)
 P for trend0.490.42
Total VTE¶
 0 cups per day26 155292.03 (1.41–2.92)Ref.Ref.Ref.
 1–2 cups per day26 548271.30 (0.89–1.89)0.60 (0.35–1.02)0.64 (0.37–1.08)0.57 (0.29–1.14)
 3–4 cups per day61 766771.35 (1.08–1.69)0.61 (0.39–0.93)0.65 (0.42–1.00)0.50 (0.28–0.89)
 5–6 cups per day72 912781.18 (0.95–1.47)0.53 (0.34–0.81)0.56 (0.36–0.87)0.60 (0.35–1.05)
 > 6 cups per day79 7231131.56 (1.29–1.87)0.77 (0.51–1.16)0.78 (0.51–1.20)0.89 (0.52–1.52)
 P for trend0.880.740.33

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure of Conflict of Interests
  7. References

Our findings indicate that moderate coffee consumption may be inversely associated with the risk of VTE. A daily coffee consumption of five to six cups was significantly associated with 33% reduced risk of VTE as compared with those who abstained from coffee. For heavy coffee drinkers (more than six cups per day), the inverse association of moderate coffee consumption was weakened. Similar risk estimates were found in separate analyses of unprovoked and provoked VTE, and in sex-stratified analyses. These findings suggest a possible U-shaped association between coffee intake and VTE.

To the best of our knowledge, only one previous observational study has investigated the association between coffee intake and VTE. In the Iowa Women’s Health study, 37 393 women aged 55–69 years were included and followed for a median of 13 years [13]. In agreement with our findings, this study reported a weak inverse association between coffee consumption and risk of VTE in a multivariable model adjusted for age, kilojoules, education level, smoking status and physical activity [13]. However, the association vanished after further adjustments for BMI and diabetes [13]. In our study, both men and women within a wide age-range were included, and the inverse association of moderate coffee consumption (five to six cups daily) and VTE remained significant after multivariable adjustments.

In contrast to the limited number of studies on coffee and VTE, the relationship between coffee consumption and arterial CVD has been extensively examined. Several prospective cohort studies have shown an inverse association between CVD and coffee consumption. The Stockholm Heart Epidemiology program [10], the Iowa Women’s Health Study [8] and a pooled analysis of data from the Health Professionals Follow-up Study and the Nurses’ Health Study [9] found a modest beneficial effect of coffee consumption on the risk of CVD mortality and overall mortality. Similar findings were reported in a Finnish cohort study [11]. In contrast, early cohort studies did not find any association between coffee consumption and CVD [15–17], whereas some case–control studies even showed a positive association between coffee consumption and the risk of non-fatal MI [18–20].

The high levels of polyphenols (especially flavanoids and phenolic acids) in coffee may explain the inverse association between coffee consumption and the risk of VTE [21]. Platelet aggregation is vital for the formation of both arterial and venous thrombi [22,23]. Dietary polyphenols inhibit platelet aggregation [22,24], and an intake of 200 mL of coffee has been shown to inhibit platelet aggregation in humans ex vivo [24]. The platelet-inhibiting effect was independent of caffeine [24]. In contrast, platelet reactivity was stimulated by caffeine administered in quantities corresponding to one cup of coffee [25]. Thus, inhibition of platelet reactivity by coffee intake may reflect the fact that the platelet-inhibiting effect of polyphenols is superior to the platelet-stimulating effect of caffeine, and thus may contribute to the protective effect of moderate coffee consumption on VTE risk.

Even though limited data are available, experimental studies suggest that coffee consumption increases fibrinolytic activity without affecting the coagulation system. Coffee ingestion has been found to shorten whole blood fibrinolysis time, an effect ascribed to caffeine [26]. Supportive results were found in another study, where caffeine exposure in humans increased tissue plasminogen activator activity and decreased plasminogen activator inhibitor levels in plasma [27]. In contrast, a study comparing the intake of decaffeinated coffee with that of caffeinated coffee among 20 human subjects found no differences in fibrinogen levels, platelet adhesiveness or clot lysis time [28]. In addition, two randomized studies comparing filtered and boiled coffee with no coffee [29], reported no differences in hemostatic variables such as fibrinogen, factor VII activity, FVIII antigen, protein C and protein S during the coffee intervention [29].

Confounding is a potential problem of cohort studies, owing to their non-randomized nature. The possible presence of personal characteristics that make the coffee abstainers different from the coffee consumers may give rise to an observed inverse association between coffee consumption and risk of VTE that is actually a result of unrecognized confounders, rather than a true association. Coffee consumption is often perceived as part of an unhealthy lifestyle [30], and when confronted with health issues, some abstain from coffee for the benefit of health [30]. Thus, the high incidence of VTE among coffee abstainers in our study could possibly be explained by underlying disease(s). However, the incidence of self-reported diabetes mellitus, cardiovascular diseases and cardiovascular risk factors such as smoking and serum lipids were not increased in coffee abstainers in our study. This, along with our finding that moderate coffee consumption tended to reduce the risk of unprovoked VTE, argues against the hypothesis that underlying diseases are confounders for the beneficial impact of coffee consumption on VTE risk.

The main strengths of our study are its prospective design, the large number of participants recruited from a general population with a high attendance rate, long follow-up time and validated VTE events. One single hospital serves the entire Tromsø population, enhancing the possibility of a complete VTE registry. The study also has limitations. Unfortunately, we did not have verified baseline information on previous history of VTE among the study subjects. Hence, some of the subjects who were treated as healthy participants during follow-up could be prevalent VTE cases who should have been excluded from the study population. However, this would lead to only a small change in the overall number of person-years at risk, and thus would presumably have a negligible influence on the risk estimates. Certain data items, such as the circumstances at the time of VTE diagnosis (e.g. provoking factors), were collected retrospectively, so the potential for ascertainment bias in some cases cannot be completely ruled out. Lifestyle habits such as coffee consumption are highly modifiable. A possible change in risk profile during follow-up may lead to misclassification bias that attenuates the real association [31]. Another possible study limitation is self-reported information. However, self-reported coffee consumption has been demonstrated to have high validity [32,33] and reproducibility over time [34,35]. Information on daily caloric intake was not available for all subjects. However, adjustments for daily caloric intake among a subgroup of participants aged 25–69 years did not alter the risk estimates. Finally, we were not able to investigate the contribution of caffeine independently, because the questionnaire did not include questions about decaffeinated coffee.

In conclusion, our findings suggest a possible U-shaped relationship between coffee consumption and VTE, and that moderate coffee consumption may be associated with reduced risk of VTE in the general population. More studies on other general population cohorts are needed to ascertain whether dietary habits such as coffee consumption are associated with the risk of VTE.

Disclosure of Conflict of Interests

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure of Conflict of Interests
  7. References

The authors state that they have no conflict of interest.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Disclosure of Conflict of Interests
  7. References