Abstract
Introduction
Non-Alcoholic Fatty Liver Disease (NAFLD) is currently the most common form of liver disease worldwide affecting all ages and ethnic groups and it has become a consistent threat even in young people. Our aim was to estimate the effect of a Low Glycemic Index Mediterranean Diet (LGIMD) on the NAFLD score as measured by a Liver Ultrasonography (LUS).
Design
NUTRIzione in EPAtologia (NUTRIEPA) is a population-based Double-Blind RCT. Data were collected in 2011 and analyzed in 2013-14.
Setting/participants
98 men and women coming from Putignano (Puglia, Southern Italy) were drawn from a previous randomly sampled population-based study and identified as having moderate or severe NAFLD.
Intervention
The intervention strategy was the assignment of a LGIMD or a control diet.
Outcome measures
The main outcome measure was NAFLD score, defined by LUS.
Results
After randomization, 50 subjects were assigned to a LGIMD and 48 to a control diet. The study lasted six months and all participants were subject to monthly controls/checks. Adherence to the LGIMD as measured by Mediterranean Adequacy Index (MAI) showed a median of 10.1. A negative interaction between time and LGIMD on the NAFLD score (-4.14, 95% CI -6.78,-1.49) was observed, and became more evident at the sixth month (-4.43, 95%CI -7.15, -1.71). A positive effect of the interaction among LGIMD, time and age (Third month: 0.07, 95% CI 0.02, 0.12; Sixth month: 0.08, 95% CI 0.03,0.13) was also observed.
Conclusions
LGIMD was found to decrease the NAFLD score in a relatively short time. Encouraging those subjects who do not seek medical attention but still have NAFLD to follow a LGIMD and other life-style interventions, may reduce the degree of severity of the disease. Dietary intervention of this kind, could also form the cornerstone of primary prevention of Type 2 Diabetes Mellitus (T2DM) and cardiovascular disease.
Key words: Hepatic steatosis, Mediterranean diet, low glycemic index Mediterranean diet
Introduction
NAFLD consists of a broad spectrum of liver diseases (1). NAFLD is currently the most common form of liver disease worldwide affecting all ages and ethnic groups. On a global level, prevalence varies between 20% and 33%. In northern Italy, prevalence of NAFLD was found to be approximately 25% in the general population, and is associated with most features of metabolic syndrome (2). NAFLD increases with age, BMI and is common in Type 2 Diabetes Mellitus (T2DM) (3). Rising prevalence of obesity and T2DM (particularly in younger people) makes NAFLD a growing threat at both the clinical and the population level (4).
Excessive hepatic fat deposition is a hallmark of NAFLD. It is linked to insulin resistance (5) and is a risk factor for T2DM and cardiovascular disease.
Although, it is predicted that some forms of NAFLD will be the leading cause of liver transplantation in the US by 2020, the optimal treatment of NAFLD remains uncertain due to the difficulties associated with adopting changes in individual's lifestyle, adverse effect of drug therapies and selection criteria, availability, and the cost of bariatric surgery (6).
Lifestyle modification is the standard treatment for NAFLD, despite the existence of little robust evidence to support this recommendation (7). This approach encompasses dietary, exercise and other behavioral changes, due to the fact that they offer a range of health benefits.
The Mediterranean Diet (MD) is a dietary pattern that has long been associated with favorable health outcomes (8, 9) and greater adherence to MD has been reported with beneficial effect on the severity of NAFLD (10). Most of the studies however, have been conducted at the clinical level where patients are highly selected (8).
A recent trial on subjects with Metabolic Syndrome (11) has shown that an energy-unrestricted MD, supplemented with extra-virgin olive oil or nuts, resulted in a substantial reduction of the risk of cardiovascular events and general mortality. The results of this study are consistent with others about the role of MD on Metabolic Syndrome (12) and diabetes (13).
In order to estimate the prevalence of some liver diseases in this Mediterranean area of Southern Italy a cohort was assembled (14) and a NAFLD prevalence of 24.6% was found. Our hypothesis was that a MD intervention would lead to greater improvements in NAFLD in this free-living population, than a diet based on the Italian National Research Institute for Foods and Nutrition (INRAN) guidelines. The standard MD of Crete and Nicotera, reported in the Ancel Keys' Seven Countries Study, used integral flour (flour from stone mills) for bread and other carbohydrate rich food (15) and in this study a Low Glycemic Index Mediterranean Diet (LGIMD), based on that diet, was experimented. The objective of this RCT was to estimate the effect of a LGIMD on the NAFLD score as measured by Liver Ultrasonography (LUS).
Materials and methods
Study design
NutriEpa was a parallel-group randomized controlled clinical trial. The sampled population was taken from the NutriEp survey, conducted at the Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) “Saverio de Bellis” (Castellana Grotte, Italy), from July 2005 to January 2007 and its details have been published elsewhere (14). In brief, in collaboration with 12 General Practitioners (GP) working in Putignano (Puglia, Italy) and, after testing the hypothesis that the sex-age group specific mean was the same among the general population and subjects of the GP clinics (p= 0.15), a random sample was drawn (for those individuals 18 years old, and older) from the GP patients' list. 2550 subjects were invited to participate in the survey; among these subjects, 2301 (90.2%) provided their written consent (according to the Helsinki Declaration).
Participant Selection
The trial was designed and conducted by the authors, and the protocol was approved initially by the Technical- Scientific review board and then by the Ethical Institutional one. Subjects with NAFLD were identified during the NutriEp enrollment process. Eligible participants were those individuals identified as having moderate or severe NAFLD (n=203). The exclusion criteria included: 1) overt cardiovascular disease and revascularization procedures; 2) stroke; 3) clinical peripheral artery disease; 4) T2DM (current treatment with insulin or oral hypoglycemic drugs, fasting glucose >126 mg/ dl, or casual glucose >200 mg/dl; 5) more than 20 gr/daily of alcohol intake; 6) severe medical condition that may impair the person to participate in a nutritional intervention study; 7) people following a special diet or involved in a program for weight loss, or who had experienced recent weight loss and 8) inability to follow a MD for religious or other reasons. Subjects were invited to participate in the trial, and after a new assessment of the severity of NAFLD using LUS, those who agreed to participate provided, in written form, their consent for participating in the trial. The trial took place at the Laboratory of Epidemiology and Biostatistics of the IRCCS “Saverio de Bellis”, Castellana Grotte (Italy), from February to November 2011 and data were analyzed in 2013-2014. All experimental subjects participated on voluntary basis.
Randomization and Masking
Participants were randomly assigned, according to a computerized random numbers sequence, to 1 of 2 treatment groups and a one-to-one ratio was used to allocate subjects.
Blinding and equipoise were strictly maintained by emphasizing to the intervention staff and participants that each diet adhered to healthy principles. With the exception of the dietitians, investigators and staff were unaware of the subjects' diet assignment. Although each dietitian followed the participant for the duration of the trial, individual assignment was made on daily random basis. Staff members who obtained outcome measurements were not informed about diet assignment. Only one of two radiologists performed outcome measurements each day and this order was also randomly assigned. In the outcome measurements relating to the third and sixth months, radiologists were unaware of the previous measurement.
Sample Size
Sample size was estimated taken into account the repeated measurement of the outcome. From a previous study (16), the mean (SD) score of NAFLD was estimated to be 4.5 (1) and 4.0 (0.5) for the treatment and control group respectively; the type I error was fixed at 0.05 (one sided) level and statistical power to 0.9. The correlation between baseline/follow-up measurements of the outcome was set to 0.4. A sample size of n1=n2=36 was estimated, in order to obtain a 1 point reduction in NAFLD score in the LGIMD group after six months.
Exposure Measurements
On the first visit, subjects were interviewed to complete a pre-coded questionnaire regarding socio-demographic issues, medical history and potential risk factors pertaining to some liver diseases. Alcohol consumption was probed (17) and the European Prospective Investigation on Cancer (EPIC) questionnaire about dietary habits was also completed. Baseline physical activity was measured by asking participants to wear a biaxial accelerometer (SenseWear®PRO3TM) over a period of seven days, including weekends.
On the second visit, a fasting venous blood sample was collected. Liver function tests and others biochemical serum markers were assessed using standard laboratory test techniques.
Body weight and height, as well as blood pressure, were measured in standard conditions at baseline, at the third and sixth month. At this point in time, dietician also reviewed (and completed if necessary) the EPIC questionnaire, taking the opportunity to highlight the advantages of following such diets.
During this second visit, at the beginning of February 2011, participants were randomly assigned to one of two dietary interventions: LGIMD or INRAN diet. The purpose of the study was explained in detail in a face-to-face interview. All participants received training so as to be able to properly fill in the Dietary Record. Personal advice was also provided in both groups, at the baseline visit and monthly thereafter. The dietary records were constructed using a weekly sheet, in which participants indicated all the food they had consumed during the course of the day. Portion size was determined taking into account the answer given in the EPIC questionnaire. The dietary record was controlled weekly during the first month of the trial, and monthly thereafter. Follow-up anthropometric measures were taken and Bioelectrical Impedance Analysis was performed monthly. LUS was performed at the third and sixth month, together with the collection of a fasting venous sample. The monthly follow-up visit also included a face-to-face interview with the dietician in order to assess the diet followed by the subject and to give, if needed, personal recommendations to achieve the “group assigned” goal.
Outcome Measurement
All subjects underwent LUS (Hitachi HI Vision E) testing. To obtain a semi-quantitative evaluation of fat in the liver, a scoring system was adopted (16, 18). NAFLD was then categorized as: absent (0), mild (1., 2.), moderate (3., 4.) and severe (5., 6.). If the subject met the trial inclusion criteria, he was invited to participate in the trial and convened for a second visit one week later. A sub-sample of 30 subjects (ten subjects at enrollment, first and second follow-up respectively randomly chosen) underwent LUS by the two radiologists separately. An overall weighted Kappa of 0.9 was obtained.
Dietary Interventions
A six-month intervention period for both the study and control diets was chosen. No advice was given as to total calories to be consumed or the level of physical activity to be undertaken.
Detailed information, relating to the LGIMD intervention and control (recommended by the World Health Organization and followed by INRAN (19), can be found in the Supplemental on line data.. Foods in LGIMD have all a low Glycemic Index (GI) and no more than 10% of total daily calories coming from saturated fats. The LGIMD was high in monounsatured fatty acids (MUFA) from olive oil and contained also omega- 3 polyunsatured fatty acids (ω3PUFA), from both plant and marine sources.
The recommended diets were provided in brochure format, with graphical explanations organized according to a traffic light system: with a list of foods that can be consumed frequently (green foods), sometimes (yellow foods) and never (red foods). The brochure also contained a dietary record, where participants daily indicated the code of each food consumed at breakfast, lunch, dinner and during snack time. A detailed description of both the intervention and control diets (the brochures provided to participants) are available as supplementary on line material. The Mediterranean Adequacy Index (MAI) was chosen as a relevant measure to evaluate the adherence to both the intervention and control diets (20). A median value of 7.5 with an inter-quantile range (IQR) of 5.4 was, as established by the reference Italian Mediterranean Diet, expected (20).
Statistical Analysis
The primary analysis was intention-to-treat and involved all participants who were randomly assigned. In this paper only socio-demographic, biochemical, BMI and Physical Activity (at baseline) data were considered. For descriptive purposes, age at enrollment (<40, 40-59 and ≥60 years old), BMI (Normal, Overweight and Obesity), Glutamic pyruvic and Glutamic oxaloacetic transaminases (altered ≥40 U/l), Triglycerides (altered ≥165 mg/dl), Cholesterol (altered ≥200 mg/dl for men and ≥220 mg/dl for women), γ-glutamyl transpeptidase (altered >25 UI/l for men and >14 UI/l for women), Insulin (altered >29 μUI/ml), Glucose (altered ≥ 127 mg/dl) and physical activity (Low <3 METs, Moderate 3-6 METs and High >6 METs) were categorized. Using age at enrollment, number of households, education, employment position, civil status and whether the individual was retired, Factor Analysis was performed to create a composite indicator of SES (21). The resulting scale (index of socio-economic position) was then standardized, reversed (higher values correspond to higher socio-economic position) and categorized as low (<0.25 quantile), medium (0.25-0.74 quantile) and high ≥ 0.75 quantile.
Cross-tabulations between interventions and socio-demographic, life-style and biological variables were performed and proportion differences were assessed using a χ2 test.
Dietary Records were analyzed using the MètaDieta® software and the results expressed as a percentage of total calorie intake for each food item consumed. In order to calculate the MAI, the four central months of the trial were considered; the first, second, third, and fourth week of the second, third, fourth, and fifth month respectively were chosen. MAI was calculated according to age-class, gender, and month to clearly describe the adherence during the trial. Furthermore, the Fatty Liver Index (FLI) (22) was estimated and its statistical analysis was performed by using a Wilcoxon matched-pairs signed-ranks test.
A Generalized Estimating Equation (GEE) (23) was performed to estimate the effect of a LGIMD and control diets on the NAFLD score (outcome). GEE models are useful in biomedical studies to estimate how the average of an outcome changes with covariates, allowing correlated response data (repeated measurements on each subject). A gamma distribution (link identity) for the response was assumed and an unstructured correlation matrix was set to the data. Gender (categorical), BMI, Liver Enzymes, Cholesterol, Triglycerides, Insulin, Glucose, physical activity level (MET/min/week-1) and age (continuous variables) were included as covariates. The results obtained are expressed in natural scale as mean ± 95% Confidence Interval (95% CI). Diet Marginal Average effects were estimated in order to illustrate intervention through age.
The statistical analysis was carried out using the statistical software, Stata version 12.1.
Results
During the NutriEp study, 203 participants were identified as having moderate or severe NAFLD; 174 of these individuals responded to our letter to take part in the trial. Of these 174, 138 had preserved the grade of severity of NAFLD (Figure 1).
Figure 1.
NutriEpa Flow Diagram. Putignano (BA), Italy, 2011
After completion of the second visit, 98 subjects agreed to participate in the trial, and were randomly selected for either the LGIMD group (n=50) or the control diet one (n=48). 6 individuals were subsequently lost in the follow-up in the intervention group and 2 in a control group. The sociodemographic characteristics of the subjects who participated in the trial are reported in Table 1. There were more men than women reflecting the distribution of NAFLD in this population. No other significant differences were found and all subject characteristics were perfectly balanced. The biochemical characteristics, BMI, FLI and NAFLD score of the participants, both at the baseline and after six months, are given in Table 2. FLI median and inter quantile range were similar at enrollment, whereas at six months they were statistically significant lower, especially for the LGIMD group. Lower levels of SGPT, Triglycerides and Glycemia were found in both groups after six months. Lower levels of Gamma-GT and HDL-Cholesterol were also observed, but only in the LGIMD group.
Table 1.
Baseline Characteristics of NutriEpa Study Participants. Putignano (BA), Italy, 2011
| DIET |
||
|---|---|---|
| CONTROL | LGIMD | |
| Age category (years)* | No (%) | No (%) |
| < 30 | 4 (100.0) | 0 (0.0) |
| 30-39 | 6 (42.9) | 8 (57.1) |
| 40-49 | 6 (35.3) | 11 (64.7) |
| 50-59 | 14 (42.4) | 19 (57.6) |
| 60-69 | 15 (65.2) | 8 (34.8) |
| 70-79 | 3 (42.9) | 4 (57.1) |
| Gender* | ||
| Male | 38 (52.8) | 34 (47.2) |
| Female | 10 (38.5) | 16 (61.5) |
| Education* | ||
| Illiterate | 0 (0.0) | 3 (100.0) |
| Elementary School | 15 (57.7) | 11 (42.3) |
| Middle School | 12 (4404) | 15 (55.6) |
| High School | 20 (52.6) | 18 (47.4) |
| College | 1 (25.0) | 3 (75.0) |
| Job* | ||
| Farmer | 1 (33.3) | 2 (66.7) |
| Worker | 5 (55.6) | 4 (44.4) |
| Employee | 21 (51.2) | 20 (48.8) |
| Trader | 1 (50.0) | 1 (50.0) |
| Freelance | 4 (50.0) | 4 (50.0) |
| Housewife | 2 (25.0) | 6 (75.0) |
| Craftsman | 14 (51.9) | 13 (48.1) |
| SEP* | ||
| Low | 2 (50.0) | 2 (50.0) |
| Medium | 38 (51.4) | 36 (48.6) |
| High | 4 (44.4) | 5 (55.6) |
| Status* | ||
| Single | 4 (57.1) | 3 (42.9) |
| Married | 42 (47.7) | 46 (52.3) |
| Widowed | 2 (66.7) | 1 (33.3) |
| Body Mass Index | ||
| Normal | 0 (0.0) | 3 (6.0) |
| Overweight | 13 (27.1) | 13 (26.0) |
| Obesity | 35 (72.9) | 34 (68.0) |
| NAFLD | ||
| Absent | 0 (0.0) | 0 (0.0) |
| Mild | 0 (0.0) | 0 (0.0) |
| Moderate | 34 (70.8) | 35 (70.0) |
| Severe | 14 (29.2) | 15 (30.0) |
| Total |
48 (49.0) |
50 (51.0) |
Pearson χ2 p>0.10. LGfMD: Low Glycemίc fndex Medίterranean Dίet; SEP: Socίo-Economic Position.
Table 2.
Biochemical Characteristics of the Participants at Baseline and Sixth Month. NutriEpa. Putignano (BA), Italy, 2011
| Baseline Diet |
Sixth Month Diet |
|||
|---|---|---|---|---|
| Control* No. (%) | LGIMD* No. (%) | Control# No. (%) | LGIMD# No. (%) | |
| Body Mass Index | ||||
| Normal | 0 (0.0) | 3 (6.0) | 0 (0.0) | 5 (10.3) |
| Overweight | 13 (27.1) | 13 (26.0) | 20 (44.2) | 19 (43.6) |
| Obesity | 35 (72.9) | 34 (68.0) | 26 (55.8) | 20 (46.2) |
| NAFLD | ||||
| Absent | 0 (0.0) | 0 (0.0) | 4 (9.3) | 4 (10.3) |
| Mild | 0 (0.0) | 0 (0.0) | 8 (16.3) | 14 (30.8) |
| Moderate | 34 (70.8) | 35 (70.0) | 27 (58.1) | 24 (53.8) |
| Severe | 14 (29.2) | 15 (30.0) | 7 (16.3) | 2 (5.1) |
| FLIH | 82.28(70.31-90.38) | 81.19(49.19-90.24) | 72.69(49.59-83.94) | 57.72(27.33-73.14) |
| SGOT | ||||
| Normal | 48 (100.0) | 50 (100.0) | 43 (100.0) | 39 (100.0) |
| Altered | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| SGPT | ||||
| Normal | 44 (91.7) | 44 (88.0) | 45 (97.7) | 39 (100.0) |
| Altered | 4 (8.3) | 6 (12.0) | 1 (2.3) | 0 (0.0) |
| Gamma-GT | ||||
| Normal | 41 (85.4) | 37 (74.0) | 40 (86.0) | 36 (82.1) |
| Altered | 7 (14.6) | 13 (26.0) | 6 (14.0) | 8 (17.9) |
| Cholesterol | ||||
| Normal | 25 (52.1) | 29 (58.0) | 24 (51.2) | 25 (56.4) |
| Altered | 23 (47.9) | 21 (42.0) | 22 (48.8) | 19 (43.6) |
| HDL-Cholesterol | ||||
| Normal | 47 (97.1) | 45 (90.0) | 44 (95.3) | 43 (97.4) |
| Altered | 1 (2.1) | 5 (10.0) | 2 (4.7) | 1 (2.6) |
| Triglycerides | ||||
| Normal | 24 (50.0) | 22 (44.0) | 33 (72.1) | 27 (61.5) |
| Altered | 14 (50.0) | 28 (56.0) | 13 (27.9) | 17 (38.5) |
| Insulin | ||||
| Normal | 44 (91.7) | 42 (84.0) | 37 (81.4) | 30 (69.2) |
| Altered | 4 (8.3) | 8 (16.0) | 9 (18.6) | 14 (30.8) |
| Glycemia | ||||
| Normal | 40 (83.3) | 36 (72.0) | 46 (100.0) | 44 (100.0) |
| Altered | 8 (16.7) | 14 (28.0) | 0 (0.0) | 0 (0.0) |
| Total |
48 (100.0) |
50 (100.0) |
46 (100.0) |
44 (100.0) |
Abbreviations: FLI: Fatty Liver Index; SGOT: Serum Glutamic-Oxalacetic Transaminase; SGPT: Serum Glutamic-PyruvateTransaminase; Gamma-GT: Gamma-Glutamyl Transferase;
Pearson χ2 p<0.05. Medίan (fnter Quantίle Range)
MAI by age-class, gender and month is shown in Table 3. Adherence to the LGIMD resulted in a median MAI of 10.1 (IQR 7.0). Adherence was found to be lower amongst women (8.4), than among men (10.5), to increase with age for men, but was found to be higher in younger women. Adherence to the control diet showed a median MAI of 4.8 (IQR 6.7), was lower amongst males (4.8) than females (8.4). Younger males (<40 years old) and younger and oldest females (<40 years old and >60 years old), showed a MAI of 10.1, 13.0 and 10.1 respectively.
Table 3.
Mediterranean Adequacy Index by Gender, Age and Month of the NutriEpa Trial
| Second Month |
Third Month |
Fourth Month |
Fifth Month |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Percentiles | |||||||||||||
| Males Age (years) | Diet | 25 | Median | 75 | 25 | Median | 75 | 25 | Median | 75 | 25 | Median | 75 |
| < 40 | LGIMD | 10.1 | 10.1 | 10.1 | 10.1 | 10.1 | 10.1 | 3.9 | 10.1 | 10.1 | 3.9 | 10.1 | 10.1 |
| CTR | 2.4 | 8.5 | 10.4 | 2.4 | 4.9 | 8.5 | 2.4 | 4.9 | 8.5 | 2.4 | 4.9 | 8.5 | |
| 40-59 | LGIMD | 7.4 | 11.8 | 17.2 | 7.4 | 11.8 | 17.2 | 6.8 | 11.0 | 16.0 | 6.8 | 11.0 | 16.0 |
| CTR | 2.9 | 8.6 | 9.9 | 4.8 | 6.3 | 9.9 | 2.9 | 7.8 | 9.9 | 2.9 | 6.3 | 9.9 | |
| 60 > | LGIMD | 12.5 | 21.1 | 45.1 | 12.5 | 21.1 | 66.6 | 12.5 | 21.1 | 66.6 | 12.5 | 12.5 | 21.1 |
| CTR | 1.5 | 3.3 | 6.4 | 1.5 | 3.3 | 6.4 | 2.1 | 3.3 | 6.4 | 2.1 | 3.3 | 6.4 | |
| Females Age (years) | |||||||||||||
| < 40 | LGIMD | 7.6 | 13.0 | 13.0 | 13.0 | 13.0 | 13.0 | 13.0 | 13.0 | 13.0 | 13.0 | 13.0 | 13.0 |
| CTR | 8.4 | 8.4 | 8.4 | 8.4 | 8.4 | 8.4 | 8.4 | 8.4 | 8.4 | 8.4 | 8.4 | 8.4 | |
| 40-59 | LGIMD | 7.0 | 8.5 | 9.5 | 7.0 | 8.5 | 10.1 | 7.0 | 8.5 | 10.1 | 7.0 | 8.5 | 10.1 |
| CTR | 0.9 | 5.7 | 5.7 | 0.2 | 0.9 | 5.7 | 0.2 | 5.7 | 5.7 | 0.9 | 0.9 | 5.7 | |
| 60 > | LGIMD | 4.5 | 5.3 | 13.8 | 4.5 | 5.3 | 13.8 | 4.5 | 5.3 | 13.8 | 4.5 | 5.3 | 13.8 |
| CTR |
8.9 |
10.1 |
15.1 |
8.9 |
10.1 |
15.1 |
8.9 |
10.1 |
15.1 |
8.9 |
10.1 |
15.1 |
|
Abbreviations: LGIMD, Low Glycemic Index Mediterranean Diet; CTR, Control.
The GEE analysis results are presented in Table 4. Significant interaction between the effect of LGIMD and time on the NAFLD score (Figure 2) was observed at both the third (-4.53, 95% CI —7.28, -1.78) and sixth month (-4.64, 95% CI -7.30, -1.99) of follow up. Also, interaction among LGIMD, time and age was significant. Predictive margins of diet vs Age are showed in Figure 2. Related to control diet all marginal average were lower (p < 0.05) up to 55 years old. After this cut point no diet effect was observed.
Table 4.
Effect of Mediterranean Diet on Stages of Non-alcoholic Fatty Liver Disease. NutriEpa, Putignano (BA), Italy 2011
| Coefficient | 95%CI1 | ||
|---|---|---|---|
| LGIMD2 | 0.94 | -1.34 | 3.21 |
| Time (3rd Month) | 1.21 | -0.75 | 3.16 |
| Time (6th Month) | -0.81 | -2.85 | 1.23 |
| LGIMD*Time | |||
| (3rd Month) | -4.14** | -6.78 | -1.49 |
| LGIMD*Time | |||
| (6th Month) | -4.43** | -7.15 | -1.71 |
| Age | 0.01 | -0.02 | 0.04 |
| LGIMD*Age | -0.02 | -0.06 | 0.02 |
| Time (3rd Month)* Age | -0.03 | -0.06 | 0.01 |
| Time (6th Month)*Age | 0.00 | -0.03 | 0.04 |
| LIGMD*Time (3rd Month)* Age | 0.07** | 0.02 | 0.12 |
| LIGMD*Time (6th Month)* Age | 0.08** | 0.03 | 0.13 |
| Gender (Female) |
0.30 |
-0.09 |
0.70 |
95% Confidence Interval;
Low Glycemic Index Mediterranean; Diet;
p< 0.01;
p< 0.05
Figure 2.
Effect and Contrast of LIGMD vs Control Diet (with 95% CI) on NAFLD Score. NutriEpa. Putignano (BA), Italy, 2011
No statistically significant effect was found between gender and age, between LGIMD and age or age and time.
Discussion
In this RCT, a LGIMD, which is rich in MUFA and ω3PUFA, with no more than 10% of total daily calories deriving from saturated fats and a low GI, was associated with a more intense reduction of the NAFLD score, as measured by LUS, than a control diet following the INRAN guidelines until 55 years old.
Although NAFLD is globally the most common form of liver disease and is now a significant cause of chronic liver disease, optimal treatment remains uncertain. The high prevalence of NAFLD observed in Western countries is probably due to the concurrent epidemics of overweight/obesity and associated metabolic complications, which are all recognized risk factors for NAFLD (24, 26).
Current therapy focuses on modifying individuals' lifestyle, such as diet and physical activity and potential pharmacologic treatments (26). Usually, NAFLD management includes weight reduction and increased physical activity (27, 28).
In this RCT, an energy unrestricted LGIMD was given as concerns have been raised with regards to the difficulty of reducing and maintaining the weight reduction over the long term (29).
Our intervention diet was rich in MUFAs and ω3PUFA. Olive oil, almost the only oil consumed in Puglia (30), is the major source of MUFAs. High MUFAs diets have found to be associated with an improvement on lipids profile, as well as glycemic control in humans (31) and animals (32). The first one includes decreased fasting plasma triacylglycerol and VLDLCholesterol concentrations without accompanying decrease in HDL (31).
PUFAs have also been shown to have an effect in NAFLD, the most commonly found being ω3PUFA. Animal studies, have shown that a ω3PUFA enriched diet increases insulin sensitivity (33), reduces intra-hepatic triglycerides (34) content and ameliorate steatohepatitis in rats (35). Evidence from cross-sectional studies show a lower consumption of ω3PUFA (36) and a higher ω6PUFA/ω3PUFA ratio in NAFLD subjects (37). Evidence from clinical trials also highlights the beneficial effect on NAFLD of a ω3PUFA supplemented diet, even in the absence of weight loss (38., 39.). To ensure an adequate ω3PUFA intake, our subjects were advised to eat purslane (Portulaca oleracea), which is eaten extensively in soups and salads in this Mediterranean Area (40).
Low-fat, low-saturated, low GI diet, as characterized by the MD of the 1960s, has showed to decrease liver fat in women (41) and middle-aged men (42) and also reduced fatty acids synthesis, after the substitution of more complex carbohydrates for sugar. It has been shown that even small decreases in liver fat maybe clinically relevant in subjects with high liver fat, (43) especially if any diet can be sustained for a longer period of time.
It is interesting that in this RCT the effect of LGIMD is found to be a decreasing trend until 55 years of age for both men and women, as it is showed in Figure 2. The “two hit model” (44) proposed for NAFLD involves excessive hepatocyte triglycerides accumulation, resulting from insulin resistance (first hit). As the first hit occurs at a young age, the MD could influence the natural history of NAFLD, by lowering the likelihood of the progression of the disease. Moreover, it has been hypothesized that diet composition may affect the severity of NAFLD by influencing the second hit. From a public health viewpoint, it is important to detect NAFLD at a relatively young age; treatment of NAFLD may form part of the primary prevention of T2DM and coronary heart disease. A recently published study has revealed for example, that the main cause of mortality amongst diabetics between 30 and 89 years old may be attributable to NAFLD (4). Meticulous assessment methods for adherence to diet must be implemented. In this trial MAI was adopted, as it is useful for individual evaluation and longitudinal studies. The median MAI found in our study was close to that of families in Crete in 1958 (44), although a higher MAI was observed amongst older men. Higher adherence to MD had been already observed in the population of southern Italy and the Greek Islands, although it has been decreasing along the time (45). The LGIMD given in this trial has shown a high adherence, probably due to an accurate follow-up of subjects and a favorable cultural background in terms of dietary habits (46).
Strengths and limitations
Some methodological issues need, however, to be considered. The strengths of this study include: the characteristics of the study subjects, who are drawn from a survey of a population sample, the statistically adequate sample size and the controlled nature of the diet intervention. Although a few subjects were lost in the follow-up stages, the small numbers involved do not give rise for concern. An intentionto- treat analysis was applied, thus there is no reason to assume that non-adherence to the protocol is related to prognosis; this RCT provides therefore an unbiased assessment of treatment efficacy (47). To control for the possible presence of residual confounding effects several covariates were included in the GEE model in an attempt to obtain more precise and valid estimates. The GEE approach estimates coefficients of the covariates considering the correlated response. Also predictive marginal averages, after adjustment, can be obtained by using this strategy which might be useful to describe mean differences along covariates of interest. Diagnosis of NAFLD was performed by LUS which has generally considerable sensitivity and specificity (48, 49), but may fail to detect hepatic fat content <25-30% (50), thus underestimating the actual liver fat infiltration, the effect of this non-differential misclassification (due to the randomization and blinding of the operator) could, however, only produce a bias toward the null (51).
Conclusions
This study (NutriEpa) compares a LGIMD with the standard national recommended diet and illustrates that a LGIMD is more effective than the standard diet in reducing NAFLD scores in subjects who do not seek medical attention. Other relevant findings include that changing dietary quality composition is a more realistic alternative to energy restricted diets and that dietary intervention at a relative young age may, as recently published, be part of primary prevention of T2DM and coronary heart disease (49).
Conflict of interest statement
ARO has no conflict of interest to declare. GM has no conflict of interest to declare. DMdP has no conflict of interest to declare. DVC has no conflict of interest to declare. CB has no conflict of interest to declare. IF has no conflict of interest to declare. MRN has no conflict of interest to declare. MC has no conflict of interest to declare. DIA has no conflict of interest to declare. AM has no conflict of interest to declare. MT has no conflict of interest to declare. MGC has no conflict of interest to declare. MC has no conflict of interest to declare. RR has no conflict of interest to declare. RI has no conflict of interest to declare. AMC has no conflict of interest to declare. All authors declare that this study was funded by a research grant from the Italian Ministry of Health. Role of the funding source: none.
Financial disclosure
All authors have no financial disclosures.
Ethical Standards
All experimental procedures were conducted in accordance with the guidelines in the Declaration of Helsinski and approved by the Ethics Committee of the IRCCS Saverio de Bellis.
Trial registration
This study is registered at www.clinicaltrials.gov NCTO1798719.
References
- 1.Loria P, Adinolfi LE, Bellentani S, et al. Practice guidelines for the diagnosis and management of nonalcoholic fatty liver disease. A Decalogue from the Italian Association for the Study of the Liver (AISF) Expert Committee. Dig Liver Dis. 2010;42:272–282. doi: 10.1016/j.dld.2010.01.021. [DOI] [PubMed] [Google Scholar]
- 2.Bedogni G, Miglioli L, Masutti F, et al. Prevalence of and risk factors for nonalcoholic fatty liver disease: the Dionysos nutrition and liver study. Hepatology. 2005;42:44–52. doi: 10.1002/hep.20734. 10.1002/hep.20734 PubMed PMID: 15895401. [DOI] [PubMed] [Google Scholar]
- 3.Kotronen A, Juurinen L, Hakkarainen A, et al. Liver fat is increased in type 2 diabetes patients and underestimated by serum alanine aminotransferase compared with equally obese nondiabetic subjects. Diabetes Care. 2008;31:165–169. doi: 10.2337/dc07-1463. 10.2337/dc07-1463 PubMed PMID: 17934148. [DOI] [PubMed] [Google Scholar]
- 4.Zoppini G, Fedeli U, Gennaro N, et al. Mortality from chronic liver diseases in diabetes. Am J Gastroenterol. 2014;109:1020–1025. doi: 10.1038/ajg.2014.132. 10.1038/ajg.2014.132 PubMed PMID: 24890439. [DOI] [PubMed] [Google Scholar]
- 5.Angelico F, Del Ben M, Conti R, et al. Insulin resistance, the metabolic syndrome, and nonalcoholic fatty liver disease. J Clin Endocrinol Metab. 2005;90:1578–1582. doi: 10.1210/jc.2004-1024. 10.1210/jc.2004-1024 PubMed PMID: 15598693. [DOI] [PubMed] [Google Scholar]
- 6.Charlton MR, Burns JM, Pedersen RA, et al. Frequency and outcomes of liver transplantation for nonalcoholic steatohepatitis in the United States. Gastroenterology. 2011;141:1249–1253. doi: 10.1053/j.gastro.2011.06.061. 10.1053/j.gastro.2011.06.061 PubMed PMID: 21726509. [DOI] [PubMed] [Google Scholar]
- 7.Mahady SE, George J. Management of Nonalcoholic Steatohepatitis. Clin Liver Dis. 2012;16:631–645. doi: 10.1016/j.cld.2012.05.003. 10.1016/j.cld.2012.05.003 PubMed PMID: 22824485. [DOI] [PubMed] [Google Scholar]
- 8.Sofi F, Abbate R, Gensini GF, et al. Accruing evidence on benefits of adherence to the Mediterranean diet on health: an updated systematic review and meta-analysis. Am J Clin Nutr. 2010;92:1189–1196. doi: 10.3945/ajcn.2010.29673. 10.3945/ajcn.2010.29673 PubMed PMID: 20810976. [DOI] [PubMed] [Google Scholar]
- 9.Galbete C, Toledo E, Toledo JB, et al. Mediterranean diet and cognitive function: the SUN project. J Nutr Health Aging. 2015;19:305–312. doi: 10.1007/s12603-015-0441-z. 10.1007/s12603-015-0441-z PubMed PMID: 25732216. [DOI] [PubMed] [Google Scholar]
- 10.Kontogianni MD, Tileli N, Margariti A, et al. Adherence to the Mediterranean diet is associated with the severity of non-alcoholic fatty liver disease. Clin Nutr. 2014;33:678–683. doi: 10.1016/j.clnu.2013.08.014. 10.1016/j.clnu.2013.08.014 PubMed PMID: 24064253. [DOI] [PubMed] [Google Scholar]
- 11.Estruch R, Ros E S-S J, et al. Primary prevention of cardiovascular disease with a Mediterranean diet. N Engl J Med. 2013;368:1279–1290. doi: 10.1056/NEJMoa1200303. 10.1056/NEJMoa1200303 PubMed PMID: 23432189. [DOI] [PubMed] [Google Scholar]
- 12.Kastorini CM, Milionis HJ, Esposito K, et al. The effect of Mediterranean diet on metabolic syndrome and its components: a meta-analysis of 50 studies and 534,906 individuals. J Am Coll Cardiol. 2011;57:1299–1313. doi: 10.1016/j.jacc.2010.09.073. 10.1016/j.jacc.2010.09.073 PubMed PMID: 21392646. [DOI] [PubMed] [Google Scholar]
- 13.Salas-Salvadó J, Bulló M, Babio N, et al. Reduction in the incidence of type 2 diabetes with the Mediterranean diet: results of the PREDIMED-Reus nutrition intervention randomized trial. Diabetes Care. 2011;34:14–19. doi: 10.2337/dc10-1288. 10.2337/dc10-1288 PubMed PMID: 20929998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cozzolongo R, Osella AR, Elba S, et al. Epidemiology of HCV infection in the general population: a survey in a southern Italian town. Am J Gastroenterol. 2009;104:2740–2746. doi: 10.1038/ajg.2009.428. 10.1038/ajg.2009.428 PubMed PMID: 19638964. [DOI] [PubMed] [Google Scholar]
- 15.Nestle M. Mediterranean diets: historical and research overview. Am J Clin Nutr. 1995;61:S1313–S1320. doi: 10.1093/ajcn/61.6.1313S. [DOI] [PubMed] [Google Scholar]
- 16.Chiloiro M, Caruso MG, Cisternino AM, et al. Ultrasound evaluation and correlates of fatty liver disease: A population Study in a Mediterranean Area. Metab Syndr Relat Disord. 2013;11:349–358. doi: 10.1089/met.2012.0169. 10.1089/met.2012.0169 PubMed PMID: 23758075. [DOI] [PubMed] [Google Scholar]
- 17.Skinner HA, Sheu WJ. Reliability of alcohol use indices. The Lifetime Drinking History and the MAST. J Stud Alcohol. 1982;43:1157–1170. doi: 10.15288/jsa.1982.43.1157. PubMed PMID: 7182675. [DOI] [PubMed] [Google Scholar]
- 18.Chiloiro M, Misciagna G. In: Handbook of Anthropometry, Physical Measures of Human Form in Health and Disease. Preedy VR, editor. Springer; New York: 2012. Ultrasonographic Anthropometry: An Application to the Measurement of Liver and Abdominal Fat; pp. 2227–2242. [Google Scholar]
- 19.INRAN (2003). Linee Guida per una sana alimentazione italiana. http://nut.entecra.it/649/Introduzione.html (accessed November 2013).
- 20.Alberti-Fidanza A, Fidanza F. Mediterranean Adequacy Index of Italian diets. Public Health Nutrition. 2004;7:937–941. doi: 10.1079/phn2004557. 10.1079/PHN2004557 PubMed PMID: 15482621. [DOI] [PubMed] [Google Scholar]
- 21.Oakes JM, Rossi PH. The measurement of SES in health research: current practice and steps toward a new approach. Soc Sci Med. 2003;56:769–784. doi: 10.1016/s0277-9536(02)00073-4. 10.1016/S0277-9536(02)00073-4 PubMed PMID: 12560010. [DOI] [PubMed] [Google Scholar]
- 22.Bedogni G, Bellentani S, Miglioli L, et al. The Fatty Liver Index: a simple and accurate predictor of hepatic steatosis in the general population. BMC Gastroenterology. 2006;6:33. doi: 10.1186/1471-230X-6-33. 10.1186/1471-230X-6-33 PubMed PMID: 17081293, PMCID 1636651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika. 1986;73:13–22. 10.1093/biomet/73.1.13 [Google Scholar]
- 24.Hamaguchi M, Kojima T, Takeda N, et al. The metabolic syndrome as a predictor of nonalcoholic fatty liver disease. Ann Intern Med. 2005;143:722–728. doi: 10.7326/0003-4819-143-10-200511150-00009. 10.7326/0003-4819-143-10-200511150-00009 PubMed PMID: 16287793. [DOI] [PubMed] [Google Scholar]
- 26.Marchesini G, Brizi M, Morselli-Labate AM, et al. Association of nonalcoholic fatty liver disease with insulin resistance. Am J Med. 1999;107:450–455. doi: 10.1016/s0002-9343(99)00271-5. 10.1016/S0002-9343(99)00271-5 PubMed PMID: 10569299. [DOI] [PubMed] [Google Scholar]
- 27.Shah K, Stufflebam A, Hilton TN, et al. Diet and exercise interventions reduce intrahepatic fat content and improve insulin sensitivity in obese older adults. Obesity. 2009;17:2162–2168. doi: 10.1038/oby.2009.126. 10.1038/oby.2009.126 PubMed PMID: 19390517, PMCID 2793412. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Luyckx FH, Desaive C, Thiry A, et al. Liver abnormalities in severely obese subjects: effect of drastic weight loss after gastroplasty. Int J Obes Relat Metab Disord. 1998;22:222–226. doi: 10.1038/sj.ijo.0800571. 10.1038/sj.ijo.0800571 PubMed PMID: 9539189. [DOI] [PubMed] [Google Scholar]
- 29.Katan MB. Weight-loss diets for the prevention and treatment of obesity. N Engl J Med. 2009;360:923–925. doi: 10.1056/NEJMe0810291. 10.1056/NEJMe0810291 PubMed PMID: 19246365. [DOI] [PubMed] [Google Scholar]
- 30.Inguaggiato R, Ferrara P, Cisternino AM, Misciagna G, Alimenti della Dieta Mediterranea. La Dieta Mediterranea della Puglia, 2012, 151–185
- 31.Garg A. High-monounsaturated-fat diets for patients with diabetes mellitus: a metaanalysis. Am J Clin Nutr. 1998;67:577S–582S. doi: 10.1093/ajcn/67.3.577S. PubMed PMID: 9497173. [DOI] [PubMed] [Google Scholar]
- 32.Hussein O, Grosovski M, Lasri E, et al. Monounsaturated fat decreases hepatic lipid content in non-alcoholic fatty liver disease in rats. World J Gastroenterol. 2007;13:361–368. doi: 10.3748/wjg.v13.i3.361. 10.3748/wjg.v13.i3.361 PubMed PMID: 17230603, PMCID 4065889. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Storlien LH, Kraegen EW, Chisholm DJ, et al. Fish oil prevents insulin resistance induced by high-fat feeding in rats. Science. 1987;237:885–888. doi: 10.1126/science.3303333. 10.1126/science.3303333 PubMed PMID: 3303333. [DOI] [PubMed] [Google Scholar]
- 34.Levy JR, Clore JN, Stevens W. Dietary n-3 polyunsaturated fatty acids decrease hepatic triglycerides in Fischer 344 rats. Hepatology. 2004;39:608–616. doi: 10.1002/hep.20093. 10.1002/hep.20093 PubMed PMID: 14999679. [DOI] [PubMed] [Google Scholar]
- 35.Sekiya M, Yahagi N, Matsuzaka T, et al. Polyunsaturated fatty acids ameliorate hepatic steatosis in obese mice by SREBP-1 suppression. Hepatology. 2003;38:1529–1539. doi: 10.1016/j.hep.2003.09.028. 10.1016/j.hep.2003.09.028 PubMed PMID: 14647064. [DOI] [PubMed] [Google Scholar]
- 36.Cortez-Pinto H, Jesus L, Barros H, et al. How different is the dietary pattern in non-alcoholic steatohepatitis patients? Clin Nutr. 2006;25:816–823. doi: 10.1016/j.clnu.2006.01.027. 10.1016/j.clnu.2006.01.027 PubMed PMID: 16677739. [DOI] [PubMed] [Google Scholar]
- 37.Zelber-Sagi S, Nitzan-Kaluski D, Goldsmith R, et al. Long term nutritional intake and the risk for non-alcoholic fatty liver disease (NAFLD): a population based study. J Hepatol. 2007;47:711–717. doi: 10.1016/j.jhep.2007.06.020. 10.1016/j.jhep.2007.06.020 PubMed PMID: 17850914. [DOI] [PubMed] [Google Scholar]
- 38.Capanni M, Calella F, Biagini MR, et al. Prolonged n-3 polyunsaturated fatty acid supplementation ameliorates hepatic steatosis in patients with non-alcoholic fatty liver disease: a pilot study. Aliment Pharmacol Ther. 2006;23:1143–1151. doi: 10.1111/j.1365-2036.2006.02885.x. 10.1111/j.1365-2036.2006.02885.x PubMed PMID: 16611275. [DOI] [PubMed] [Google Scholar]
- 39.Tanaka N, Sano K, Horiuchi A, et al. Highly purified eicosapentaenoic acid treatment improves nonalcoholic steatohepatitis. J Clin Gastroenterol. 2008;42:413–418. doi: 10.1097/MCG.0b013e31815591aa. 10.1097/MCG.0b013e31815591aa PubMed PMID: 18277895. [DOI] [PubMed] [Google Scholar]
- 40.Omara-Alwala TR, Mebrahtu T, Prior DE, et al. Omega-three fatty acids in purslane (Portulaca oleracea) Tissues. J Am Oil Chem Soc. 1991;68:198–199. 10.1007/BF02657769 [Google Scholar]
- 41.Westerbacka J, Lammi K, Häkkinen AM, et al. Dietary fat content modifies liver fat in overweight non-diabetic subjects. J Clin Endocrinol Metab. 2005;90:2804–2809. doi: 10.1210/jc.2004-1983. 10.1210/jc.2004-1983 PubMed PMID: 15741262. [DOI] [PubMed] [Google Scholar]
- 42.van Herpen NA, Schrauwen-Hinderling VB, Schaart G, et al. Three weeks on a highfat diet increases intrahepatic lipid accumulation and decreases metabolic flexibility in healthy overweight men. J Clin Endocrinol Metab. 2011;96:E691–E695. doi: 10.1210/jc.2010-2243. 10.1210/jc.2010-2243 PubMed PMID: 21252252. [DOI] [PubMed] [Google Scholar]
- 43.Utzschneider KM, Bayer-Carter JL, Arbuckle MD, et al. Beneficial effect of a weight-stable, low-fat/low-saturated fat/low-glycemic index diet to reduce liver fat in older subjects. Br J Nutr. 2013;109:1096–1104. doi: 10.1017/S0007114512002966. 10.1017/S0007114512002966 PubMed PMID: 22849970. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Day CP & James OF. Steatohepatitis: a tale of two “hits”? Gastroenterology 1998; 114: 842–845. 10.1016/S0016-5085(98)70599-2 [DOI] [PubMed]
- 45.Alberti A, Fruttini D, Fidanza F. The Mediterranean Adequacy Index: further confirming results of validity. Nutr Metab Cardiovasc Dis. 2009;19:61–66. doi: 10.1016/j.numecd.2007.11.008. 10.1016/j.numecd.2007.11.008 PubMed PMID: 18337072. [DOI] [PubMed] [Google Scholar]
- 46.Middleton G, Keegan R, Smith MF, et al. Implementing a Mediterranean Diet Intervention into an RCT: Lessons Learned from a Non-Mediterranean Based Country. J Nutr Health Aging2015; 19: 1019–1022. 10.1007/s12603-015-0522-z. [DOI] [PubMed]
- 47.Montori VM, Guyatt GH. Intention-to-treat principle. CMAJ. 2001;165:1339–1341. PubMed PMID: 11760981, PMCID 81628. [PMC free article] [PubMed] [Google Scholar]
- 48.Joy D, Thava VR, Scott B. Diagnosis of fatty liver disease: is biopsy necessary? Eur J Gastroenterol Hepatol. 2003;15:539–543. doi: 10.1097/01.meg.0000059112.41030.2e. PubMed PMID: 12702913. [DOI] [PubMed] [Google Scholar]
- 49.Wieckowska A, McCullough AJ, Feldstein AE. Noninvasive diagnosis and monitoring of nonalcoholic steatohepatitis: present and future. Hepatology. 2007;46:582–589. doi: 10.1002/hep.21768. 10.1002/hep.21768 PubMed PMID: 17661414. [DOI] [PubMed] [Google Scholar]
- 50.Saadeh S, Younossi ZM, Remer EM, et al. The utility of radiological imaging in nonalcoholic fatty liver disease. Gastroenterology. 2002;123:745–750. doi: 10.1053/gast.2002.35354. 10.1053/gast.2002.35354 PubMed PMID: 12198701. [DOI] [PubMed] [Google Scholar]
- 51.Cameron SV, Heckman JJ. Life Cycle Schooling and Dynamic Selection Bias: Models and Evidence for Five Cohorts of American Males. The Journal of Political Economy. 1998;106:262–333. 10.1086/250010 [Google Scholar]
Uncited references
- 25.Bedogni G, Bellentani S. Fatty liver disease: how frequent is it and why? Ann Hepatol. 2004;3:63–65. PubMed PMID: 15257248. [PubMed] [Google Scholar]
- 52.Bazzano LA, Hu T, Reynoldas K, et al. Effects of Low-Carbohydrate and Low-Fat Diets: A Randomized Trial. Ann Intern Med. 2014;161:309–318. doi: 10.7326/M14-0180. 10.7326/M14-0180 PubMed PMID: 25178568, PMCID 4428290. [DOI] [PMC free article] [PubMed] [Google Scholar]