Summary
Background
Accumulating evidence supports health benefits of dietary fibre, such as improving lipid profiles, lowering blood pressure and improving insulin sensitivity, but evidence from comprehensive investigation of dietary fibre intake and mortality from cardiovascular disease (CVD) and all cancers is limited.
Aims
To quantitatively assess the association between dietary fibre intake and mortality from CVD and all cancers.
Methods
We performed a meta-analysis of prospective cohort studies. Eligible studies were identified by searching PubMed and Embase databases for all articles published up to September 2014 and via hand searching. Study-specific estimates adjusting for potential confounders were combined to calculate pooled relative risks (RRs) with 95% confidence intervals (CIs), using a random-effects model.
Results
We found 15 studies that examined the association between dietary fibre and mortality from CVD, coronary heart disease (CHD) and all cancers. The pooled RRs of CVD, CHD and all-cancer mortality for the highest versus lowest category of dietary fibre were 0.77 (95% CI: 0.71–0.84), 0.76 (95% CI: 0.67–0.87) and 0.86 (95% CI: 0.79–0.93), respectively. In a dose-response meta-analysis, the pooled RRs for an increment of 10 g/day in dietary fibre intake were 0.91 (95% CI: 0.88–0.94) for CVD, 0.89 (95% CI: 0.85–0.93) for CHD and 0.94 (95% CI: 0.91–0.97) for all cancers.
Conclusions
Our findings suggest that high dietary fibre intake is associated with a reduced risk of mortality from CVD and all cancers. These results support the current recommendation that high dietary fibre intake should be part of a healthy diet.
Résumé
Justification
Des données convergentes sont en faveur du bénéfice des fibres diététiques, qui contribuent à améliorer le profil lipidique, à réduire le niveau de pression artérielle et à améliorer la sensibilité à l’insuline, mais la preuve de cette hypothèse reste à démontrer en particulier en ce qui concerne le rôle de la prise de fibres diététiques sur le pronostic, mortalité cardiovasculaire, en particulier chez les patients souffrant d’un cancer.
Objectif
Évaluer de façon quantitative l’association entre la consommation des fibres diététiques et la mortalité cardiovasculaire et par cancer en utilisant une méthodologie standard, méta-analyse des études de cohortes prospectives.
Méthode
Des études éligibles ont été identifiées au travers d’une recherche PubMed et Embase pour tous les articles jusqu’en septembre 2014 complétée par une recherche manuelle. L’estimation a été obtenue après ajustement sur les variables de confusion, afin de calculer un risque relatif global, l’IC 95 % en utilisant un modèle aléatoire.
Résultats
Parmi les 15 études qui ont examiné l’association entre prise de fibres diététiques et mortalité cardiovasculaire, coronaire et par cancer, la valeur du risque relatif de la mortalité en comparant la consommation la plus élevée de fibres diététiques par rapport à la consommation la plus faible était respectivement de 0,77 (IC 95 % : 0,71–0,84), 0,76 (IC 95 % : 0,67–0,87) et 0,86 (IC 95 % : 0,79–0,93). Dans une méta-analyse dose-réponse, le risque relatif global pour l’augmentation des consommations de fibres diététiques de 10 g/j donne le résultat suivant pour le risque relatif : 0,91 (IC 95 % : 0,88–0,94) pour la mortalité cardiovasculaire 0,89 (IC 95 % : 0,85–0,93) pour la mortalité coronaire et 0,94 (IC 95 % : 0,91–0,97) pour la mortalité par cancer.
Conclusion
Cette méta-analyse suggère que la consommation élevée de fibres diététiques est associée à une réduction significative de la mortalité cardiovasculaire et par cancer. Ces résultats sont en faveur d’une recommandation accrue de fibres diététiques faisant partie intégrante d’un régime préservant le niveau de santé.
Background
Cardiovascular disease (CVD) and cancer are the leading causes of death worldwide . In particular, a large portion of premature deaths (death before the age of 75 years) were from CVD and cancer , so we need to develop effective preventive strategies to reduce mortality from CVD and cancer. Several modifiable factors (i.e. smoking, physical activity, body mass index [BMI] and healthy dietary pattern) have been found to be related to CVD and cancer . Dietary fibre is rich in fruits, vegetables and whole grains, and consists of portions of plant foods that are edible and non-digestible by humans . Dietary fibre intake is widely recognized as a part of a healthy diet, and higher intake is inversely associated with disease risk factors such as dyslipidaemia, obesity, hypertension and diabetes .
Accumulating evidence from epidemiological studies has shown that high intake of dietary fibre is inversely associated with the incidence of CVD (9% reduction in risk of CVD for an increase in dietary fibre of 7 g per day) . Previous research into dietary fibre intake and CVD focused on the incidence of CVD as an endpoint. Recently, a meta-analysis of prospective cohort studies showed a significant inverse association between dietary fibre intake and all-cause mortality . There is a growing body of epidemiological studies on dietary fibre intake and mortality from CVD, coronary heart disease (CHD) or all cancers , but a comprehensive assessment of dietary fibre intake and mortality from CVD, CHD and all cancers has not been carried out. Therefore, we conducted a systematic review and meta-analysis of prospective cohort studies to assess the risk of mortality from CVD, CHD and all cancers in relation to dietary fibre intake in the general population.
Methods
Literature search and study selection
We used the electronic PubMed and Embase databases to search for eligible studies published in English up to September 2014. The following key terms were used in searching: (“fibre” or “fibre”) combined with (“cardiovascular disease” or “CVD” or “coronary disease” or “CHD” or “ischaemic disease” or “IHD” or “cancer” and “mortality” or “death”). Additionally, we reviewed the reference lists of the selected articles and published reviews to identify more studies. We selected studies according to following criteria: prospective cohort design; the exposure of interest was dietary fibre intake; the outcome of interest was mortality from CVD, CHD, ischaemic heart disease (IHD) or all cancers; and RRs with 95% confidence intervals (CIs) were provided. Studies that reported relative risks (RRs) from incidence of CVD, CHD, IHD or cancer were not included. Studies conducted in patients with specific diseases were excluded.
Data extraction
Data extraction was conducted by two investigators (Y.K. and Y.J.), independently, according to the Meta-analysis of observational studies in epidemiology (MOOSE) guidelines , and any discrepancies were resolved by referencing original articles and having a further discussion. The following information was extracted from each included study: first author’s last name; year of publication; cohort name; geographical region; follow-up period; number of cases; number of subjects or person-time; adjustment factors; RR and 95% CI for the association between various levels of fibre intake and mortality from CVD, CHD, IHD or all cancers. We used the RR from multivariate models with the greatest degree of adjustment for potential confounders if the study reported several RRs on this subject.
Quality assessment of individual studies
Two investigators (Y.K. and Y.J.) independently assessed the quality of studies included in meta-analysis by applying the Newcastle-Ottawa Scale for the following aspects: selection of study groups (0–4 points); comparability of groups (0–3 points); and ascertainment of outcome (0–6 points). Disagreements between the investigators on more than one score were resolved by consensus. Studies with a total of more than 10 points (out of 13) were considered to be of high quality, and those with 7–9 points were considered to be of good quality.
Statistical analysis
To compute a pooled RR and its 95% CI for the highest versus lowest category of dietary fibre intake from original studies, we combined the natural logarithm of the RRs from each study, using the DerSimonian and Laird random-effects models, which incorporate both within- and between-study variations . For studies that had not used the lowest category as a reference, the RRs and 95% CIs were recalculated. If studies provided RRs for both total dietary fibre intake and dietary fibre intake from each food source, we included a RR for total dietary fibre intake in the main analysis. The RRs of dietary fibre intake from each food source were used in the subgroup analyses by source of fibre. The study-specific RRs as well as a pooled RR are presented as forest plots, where the size of data markers (squares) corresponds to the inverse of variance of the natural logarithm of RR from each study, and a diamond indicates a pooled RR. We assessed statistical heterogeneity among the studies using the Cochran Q statistic and inconsistency was quantified by the I 2 statistic .
To examine dose-response relationships across various categories of dietary fibre intake, we used a generalized least-squares trend estimation analysis based on the method developed by Greenland and Longnecker . After estimating study-specific slopes from the correlated natural logarithm of the RRs across dietary fibre categories in each study , we combined the generalized least-squares trend-estimated study-specific slopes with studies that reported slope estimates to derive an overall average slope. If the studies did not report mean or median levels of dietary fibre intake, we considered the midpoint of the category as a dose. The lowest and highest open-ended categories were assumed to have an equal interval of intake as the adjacent category. As two studies did not report the number of deaths or person-time across dietary fibre categories , a total of 13 studies were included in the dose-response meta-analysis for CHD mortality , CVD mortality and all-cancer mortality .
To evaluate whether the pooled RR in the main analysis was affected markedly by a single study, we carried out a sensitivity analysis, excluding one study at a time. We also performed subgroup analyses stratified by sex, geographical region (Europe/USA/others), adjustment for covariates (age, BMI, smoking, alcohol and physical activity), source of dietary fibre (cereals/vegetables/fruits/legumes) and water solubility (soluble/insoluble), and conducted meta-regression analyses to test for variations in pooled RRs among the subgroups. Finally, publication bias was evaluated through a funnel plot as well as the statistical tests of Begg and Mazumdar and Egger et al. . A two-tailed P -value < 0.05 was considered statistically significant. All statistical analyses were performed by using Stata/SE software, version 12.0 (Stata Corporation, College Station, TX, USA).
Results
Study characteristics
We identified a total of 15 studies from 12 cohorts that reported RRs of dietary fibre intake and mortality from CVD, CHD, IHD or cancers ( Fig. 1 ). The main characteristics of the studies included in meta-analysis are presented in Table 1 . Six studies provided RRs of mortality from CVD only , four studies provided RRs of mortality from CHD only , two studies provided RRs of IHD only and three studies provided RRs of mortality from both CVD and CHD . Among the 15 studies, only three studies reported RRs of mortality from all cancers . By geographical region, seven studies were conducted in Europe , five in the USA , two in Australia and one in Japan . The follow-up periods ranged from 5.9 to 40 years; the mean follow-up time was 14.5 years. The age of the subjects ranged from 20 to 85 years. Most of the studies assessed dietary fibre intake using food frequency questionnaires , while two studies used 24-hour dietary recall and one study used a cross-check dietary history method . One study from the UK assessed dietary fibre intake using both food frequency questionnaires and a 7-day food diary . All of the 15 studies adjusted for age and 13 studies adjusted for BMI or smoking . Ten studies adjusted for alcohol intake or physical activity . With regard to quality assessment, the mean quality assessment score was 10 (range 8–13). Six studies had more than 10 points, indicating high quality ; the other nine studies had 8 or 9 points, indicating good quality .
| Reference | Study location | Follow-up period (years) | Age at baseline (years) | Study size | RR (95% CI) | Adjustment factors | ||
|---|---|---|---|---|---|---|---|---|
| Subjects ( n ) | Deaths ( n ) | Fibre category (g/day) | ||||||
| Khaw and Barrett-Connor, 1987 | USA a | 12 | 50–79 | All: 859 | IHD: all: 65 | Per 6 | All: 0.79 (0.63, 0.98) | Age, systolic BP, cholesterol, fasting plasma glucose, BMI, cigarette smoking, plus oestrogen use (F only) |
| M: 356 | IHD: M: 42 | M: 0.85 (0.64, 1.11) | ||||||
| F: 503 | IHD: F: 23 | F: 0.67 (0.45, 1.00) | ||||||
| < 16 vs > 16 | M: 0.33 (0.14, 0.77) | Age | ||||||
| F: 0.37 (0.11, 1.24) | ||||||||
| Pietinen et al., 1996 | Finland b | 6.1 | 50–69 | M: 21,930 | CHD: M: 635 | 16.1 | 1.00 | Age, smoking, BMI, BP, intake of energy, alcohol and saturated fatty acids, education, physical activity, intake of beta-carotene, vitamin C and vitamin E |
| 20.7 | 0.91 (0.72–1.15) | |||||||
| 24.3 | (0.83 0.64–1.06) | |||||||
| 28.3 | (0.72 0.55–0.93) | |||||||
| 34.8 | (0.73 0.56–0.95) | |||||||
| Wolk et al., 1999 | USA c | 10 | 37–64 | F: 68,782 | CHD: F: 162 | 11.5 | 1.00 | Age |
| 14.3 | 0.83 (0.52, 1.31) | |||||||
| 16.4 | 0.74 (0.46, 1.18) | |||||||
| 18.8 | 0.73 (0.46, 1.16) | |||||||
| 22.9 | 0.41 (0.23, 0.70) | |||||||
| Bazzano et al., 2003 | USA d | 19 | 25–74 | 9248 | CHD: 668 | < 7.7 n | 1.00 | Age, alcohol intake, BMI, smoking, education level, DM, physical activity, sex, race, systolic BP, serum total cholesterol concentration, saturated fat intake, total calorie intake |
| 7.7–11.0 n | 0.87 (0.68, 1.12) | |||||||
| 11.1–15.9 n | 0.79 (0.63, 0.99) | |||||||
| > 15.9 n | 0.85 (0.65, 1.10) | |||||||
| Per 10 n | 0.92 (0.84, 1.01) | |||||||
| CVD: 1198 | < 7.7 n | 1.00 | ||||||
| 7.7–11.0 n | 0.95 (0.78, 1.15) | |||||||
| 11.1–15.9 n | 0.84 (0.72, 0.97) | |||||||
| > 15.9 n | 0.93 (0.77, 1.12) | |||||||
| Per 10 n | 0.97 (0.93, 1.02) | |||||||
| Streppel et al., 2008 | Netherlands e | 40 | 49 ± 6 | M: 1373 | CHD: M: 348 | Per 10 | 0.87 (0.71, 1.07) | Age, total intake of energy, intake of saturated fat, trans unsaturated fatty acids and cis polyunsaturated fatty acids, intake of alcohol, wine and fish, prescribed diet, number of cigarettes smoked, duration of cigarette smoking, cigar or pipe smoking, BMI, socioeconomic status |
| Kaushik et al., 2009 | Australia f | 13 | ≥ 49 | 2897 | CHD: NR | Tertile 1: 3.0 | 0.94 (0.73, 1.22) | Age, sex, systolic BP, diastolic BP, use of antihypertensives, BMI, smoking status, educational qualifications, fair or poor self-rated health, history of MI and stroke, DM, energy intake |
| Tertile 2: 6.5 | 1.22 (0.95, 1.58) | |||||||
| Tertile 3: 11.0 | 1.00 | |||||||
| Eshak et al., 2010 | Japan g | 14.3 | 40–79 | M: 23,119 | CVD: M: 1063 | Q1: M < 7.8 | M: 1.00 | Age, BMI, history of hypertension, history of DM, alcohol intake, smoking, education level, hours of exercise, hours of walking, perceived mental stress, sleep, intake of fish, saturated fatty acids, (n-3) fatty acids, sodium, folate and vitamin E |
| F: 35,611 | CVD: F: 1017 | Q1: F < 8.5 | F: 1.00 | |||||
| Q2: M: 7.8–9.4 | M: 1.06 (0.92, 1.22) | |||||||
| Q2: F: 8.5–9.9 | F: 1.21 (0.94, 1.44) | |||||||
| Q3: M: 9.5–10.8 | M: 0.92 (0.80, 1.06) | |||||||
| Q3: F: 10.0–11.1 | F: 1.06 (0.82, 1.19) | |||||||
| Q4: M: 10.9–12.6 | M: 0.86 (0.55, 1.55) | |||||||
| Q4: F: 11.2–12.7 | F: 0.85 (0.69, 0.99) | |||||||
| Q5: M > 12.6 | M: 0.83 (0.63, 1.09) | |||||||
| Q5: F > 12.7 | F: 0.82 (0.57, 0.97) | |||||||
| CHD: 231 M | Q1: M < 7.8 | M: 1.00 | ||||||
| CHD: 484 F | Q1: F < 8.5 | F: 1.00 | ||||||
| Q2: M: 7.8–9.4 | M: 0.83 (0.62, 1.12) | |||||||
| Q2: F: 8.5–9.9 | F: 1.03 (0.76, 1.48) | |||||||
| Q3: M: 9.5–10.8 | M: 0.69 (0.51, 0.93) | |||||||
| Q3: F: 10.0–11.1 | F: 0.86 (0.61, 1.13) | |||||||
| Q4: M: 10.9–12.6 | M: 0.59 (0.43, 0.81) | |||||||
| Q4: F: 11.2–12.7 | F: 0.81 (0.52, 0.99) | |||||||
| Q5: M > 12.6 | M: 0.81 (0.61, 1.09) | |||||||
| Q5: F > 12.7 | F: 0.80 (0.57, 0.97) | |||||||
| Buyken et al., 2010 | Australia f | 13 | ≥ 49 | M: 1245 | CVD: M: 151 | Tertile 1: M: 18.4 | M: 1.00 | Age, energy intake, total fibre residuals, dietary glycaemic index residuals, total fat intake, whether underweight, current smoking, use of corticosteroid drugs at baseline |
| F: 1490 | CVD: F: 109 | Tertile 1: F: 19.7 | F: 1.00 | |||||
| Tertile 2: M: 25.9 | M: 0.74 (0.49, 1.13) | |||||||
| Tertile 2: F: 24.8 | F: 0.90 (0.55, 1.46) | |||||||
| Tertile 3: M: 36.4 | M: 0.84 (0.53, 1.34) | |||||||
| Tertile 3: F: 36.2 | F: 0.88 (0.53, 1.46) | |||||||
| Baer et al., 2011 | USA c | 18 | 52.5 | 50,112 | CVD: 1026 | Per 4 | 0.82 (0.69, 0.97) | Age, BMI, weight change since age 18 years, height, smoking status, smoking amount/duration, physical activity, alcohol intake, nut consumption, polyunsaturated fat intake, glycaemic load, dietary cholesterol, systolic BP, use of antihypertensives, personal history of DM, parental MI before age 60 years, time since menopause |
| Akbaraly et al., 2011 | UK h | 18 | 49.5 (39–63) | 6926 | CVD: 141 | Per 4.5 | 0.91 (0.74, 1.12) | Age, modified total AHEI score excluding the component considered in the analysis, sex, ethnicity, occupational grade, marital status, smoking status, total energy intake, physical activity, BMI categories, prevalent CVD, type 2 DM, hypertension, dyslipidaemia, metabolic syndrome, inflammatory markers |
| Park et al., 2011 | USA i | 9 | 50–71 | M: 219,123 | CVD: M: 5248 | Q1: M: 12.6 | M: 1.00 | Age, race/ethnicity, education, marital status, health status, BMI, physical activity, smoking status, time since quitting, smoking dose, intake of alcohol, red meat and total fruit/vegetables, total energy intake |
| F: 168,999 | CVD: F: 2147 | Q1: F: 10.8 | F: 1.00 | |||||
| Q2: M: 16.4 | M: 0.89 (0.82, 0.96) | |||||||
| Q2: F: 14.3 | F: 0.85 (0.75, 0.96) | |||||||
| Q3: M: 19.4 | M: 0.82 (0.75, 0.90) | |||||||
| Q3: F: 17.0 | F: 0.78 (0.68, 0.90) | |||||||
| Q4: M: 22.9 | M: 0.80 (0.73, 0.89) | |||||||
| Q4: F: 20.1 | F: 0.72 (0.61, 0.84) | |||||||
| Q5: M: 29.4 | M: 0.76 (0.68, 0.85) | |||||||
| Q5: F: 25.8 | F: 0.66 (0.55, 0.79) | |||||||
| Cancer: M: 8244 | Q1: M: 12.6 | M: 1.00 | ||||||
| Cancer: F: 4927 | Q1: F: 10.8 | F: 1.00 | ||||||
| Q2: M: 16.4 | M: 0.98 (0.91, 1.04) | |||||||
| Q2: F: 14.3 | F: 0.93 (0.85, 1.01) | |||||||
| Q3: M: 19.4 | M: 0.94 (0.87, 1.01) | |||||||
| Q3: F: 17.0 | F: 0.88 (0.80, 0.97) | |||||||
| Q4: M: 22.9 | M: 0.87 (0.81, 0.94) | |||||||
| Q4: F: 20.1 | F: 0.88 (0.79, 0.98) | |||||||
| Q5: M: 29.4 | M: 0.83 (0.76, 0.92) | |||||||
| Q5: F: 25.8 | F: 0.96 (0.85, 1.08) | |||||||
| Crowe et al., 2012 | Europe j | 11.5 | 40–85 | 306,331 | IHD: 2381 | < 17.5 | 1.00 | Age, alcohol intake, BMI, physical activity, marital status, highest education level, current employment, hypertension, hyperlipidaemia, angina pectoris, DM, polyunsaturated to saturated fat ratio, total energy intake |
| 17.5–22.4 | 0.80 (0.71, 0.89) | |||||||
| 22.5–27.4 | 0.78 (0.69, 0.89) | |||||||
| ≥ 27.5 | 0.77 (0.66, 0.89) | |||||||
| Per 10 | 0.85 (0.73, 0.99) | |||||||
| Chuang et al., 2012 | Europe k | 12.7 | 25∼70 | M: 130,564 | CVD: M: 2489 | < 16.4 | M: 1.00 | Age, education, smoking, alcohol intake, BMI, physical activity, total energy intake, plus anytime use of menopausal hormone therapy (F only) |
| F: 322,153 | CVD: F: 2115 | F: 1.00 | ||||||
| 16.4 to < 20.1 | M: 0.91 (0.79, 1.03) | |||||||
| F: 0.83 (0.72, 0.95) | ||||||||
| 20.1 to < 23.6 | M: 0.82 (0.71, 0.94) | |||||||
| F: 0.80 (0.69, 0.93) | ||||||||
| 23.6 to < 28.5 | M: 0.81 (0.70, 0.94) | |||||||
| F: 0.73 (0.61, 0.86) | ||||||||
| ≥ 28.5 | M: 0.83 (0.71, 0.98) | |||||||
| F: 0.67 (0.55, 0.82) | ||||||||
| Per 10 | M: 0.84 (0.74, 0.94) | |||||||
| F: 0.79 (0.67, 0.94) | ||||||||
| Cancer: M: 4039 | < 16.4 | M: 1.00 | ||||||
| Cancer: F: 5575 | F: 1.00 | |||||||
| 16.4 to < 20.1 | M: 0.93 (0.84, 1.03) | |||||||
| F: 0.88 (0.81, 0.95) | ||||||||
| 20.1 to < 23.6 | M: 0.84 (0.75, 0.93) | |||||||
| F: 0.92 (0.84, 1.00) | ||||||||
| 23.6 to < 28.5 | M: 0.92 (0.82, 1.03) | |||||||
| F: 0.90 (0.81, 0.99) | ||||||||
| ≥ 28.5 | M: 0.82 (0.72, 0.93) | |||||||
| F: 0.82 (0.73, 0.92) | ||||||||
| Per 10 | M: 0.91 (0.84, 0.99) | |||||||
| F: 0.86 (0.78, 0.95) | ||||||||
| Threapleton et al., 2013 | UK l | 14.3 | 51.8 | F: 31,036 | CHD: F: 113 | Q1: 21.0 | 1.00 | Age, BMI, calories from carbohydrates, fat and protein, ethanol intake, METs (1 kcal/kg/h), smoking status, socioeconomic status |
| Q2: 30.0 | 1.18 (0.68, 2.03) | |||||||
| Q3: 36.8 | 0.82 (0.43, 1.58) | |||||||
| Q4: 44.8 | 0.99 (0.52, 1.86) | |||||||
| Q5: 63.0 | 0.72 (0.28, 1.83) | |||||||
| Per 11 | 0.96 (0.73, 1.26) | |||||||
| CVD: F: 230 | Q1: 21.0 | 1.00 | ||||||
| Q2: 30.0 | 0.84 (0.56, 1.25) | |||||||
| Q3: 36.8 | 0.62 (0.39, 0.98) | |||||||
| Q4: 44.8 | 0.72 (0.46, 1.13) | |||||||
| Q5: 63.0 | 0.61 (0.32, 1.17) | |||||||
| Per 11 | 0.91 (0.76, 1.08) | |||||||
| Buil-Cosiales et al., 2014 | Spain m | 5.9 years | M (55–75) | 7216 | CVD: 103 | Q1: 17 | 1.00 | Age, sex, smoking status, DM, BMI, baseline systolic and diastolic arterial BP, intervention group use of statins, alcohol intake, educational level, physical activity, total energy intake, plus stratification by recruitment centre |
| F (60–75) | Q2: 21 | 0.55 (0.31, 0.95) | ||||||
| Q3: 24 | 0.57 (0.32, 1.02) | |||||||
| Q4: 28 | 0.65 (0.35, 1.19) | |||||||
| Q5: 35 | 0.46 (0.23, 0.93) | |||||||
| Cancer: 169 | Q1: 17 | 1.00 | ||||||
| Q2: 21 | 0.67 (0.42, 1.06) | |||||||
| Q3: 24 | 0.78 (0.50, 1.23) | |||||||
| Q4: 28 | 0.59 (0.35, 0.98) | |||||||
| Q5: 35 | 0.69 (0.42, 1.14) | |||||||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
