Comparative Analysis Of Low Fat And Low Carbohydrate Diets On Weight Reduction And Insulin Secretion

Aim of the Study

The authors of this study conducted a comparative analysis on the effects of weight reduction imparted by consumption of low carbohydrate and a low fat diet. The authors also assessed the insulin secretion and genotype patterns of the subjects concerned with consumption of the respective diets (Gardner et al., 2018).

The aim was associated with the determination of effects of consumption of a healthy low fat diet, compared with the effects associated with the consumption of a healthy low carbohydrate on the loss of weight in individuals who were obese. The study also aimed to evaluate the association of insulin secretion and genotype patterns of the subjects with the effects of weight loss resulting from the consumption of low fat and low carbohydrate diets (Gardner et al., 2018).

The primary hypothesis stated a statistically significant interaction between genotype pattern and diet for weight reduction. The second hypothesis stated  a statistically significant interaction between diet and insulin secretion for weight loss (Gardner et al., 2018).

Participants

The authors recruited 600 adults from the Bay areas of California, via emails and mass media advertisements from previously conducted nutritional based studies. The participant group consisted of men and women belonging to the premenopausal stage, within the age group of 18 to 50 years and within a Body Mass Index of 28 to 40. Individuals who were suffering from metabolic disorders such as uncontrolled diabetes, hypertension, cancer, renal or hepatic disorders or in the stage of lactation or pregnancy were excluded. Individuals who were consuming medications such as antihypertensive, hypoglycemic, psychiatric or lipid lowering drugs were excluded. Individuals who remained stable for 3 months before collection of baseline data despite taking medications were included (Gardner et al., 2018).

The authors conducted a randomized controlled trial using allocation sequence generation procedure from a computer. The low fat diet group were prescribed a diet with reduced quantities of nuts, cooking oils, fatty cuts of meats and full fat sources of dairy. The low carbohydrate diet group were instructed to follow a diet with reduced quantities of starchy vegetables, whole grains and legumes. The energy expenditures were measured using Stanford Seven Day Activity Recall Questionnaire and dual energy X-ray absorptiometry was utilized for the purpose of measurement of body composition. For the evaluation of genotype patterns and insulin secretions, the UK Biobank Axiom microarray was used (Gardner et al., 2018).

The selected participants were instructed to undertake a training session of one month, where they were educated concerning the diet principles to be followed. Behavior modification and emotional awareness were used to encourage participant adherence. In order to test the hypothesis of diet and insulin secretion was measured with the aid of a Wald test, as a 2 way interaction for 12 months between diet and time point. For testing the hypothesis pertaining to genotype pattern (of baseline INS-30), a Wald test was used by the authors as a 3 way interaction between diet, time and genotype (Gardner et al., 2018).

Study Design

Main Findings

It was uncovered that 180 subjects possessed a low carbohydrate genotype and 244 subjects possessed a low fat genotype, with a mean baseline of INS-30, 93 µIU/ml. About 481 subjects were able to complete the trial. When compared between healthy low carbohydrate and healthy low fat diets, the distribution of macronutrients for a period of 12 months of undertaking the dietary interventions were, 30% vs. 48% for carbohydrates, 45% vs. 29% for fat and 23% vs. 21% for proteins. The authors of the study recorded the changes in weight where the results exhibited a – 6.0 kg loss in healthy low carbohydrate diet as compared to the – 5.3 kg loss in weight in the participants consuming healthy low fat diet (Mean values of differences in weight: 0.7 kg, 95% Confidence Interval, -0.2 to 1.6 kg). No statistically significant interactions were recorded pertaining to diet-insulin secretion (INS-30) or diet-genotype pattern. (p = 0.20) (Gardner et al., 2018).

The authors clearly stated a primary and a secondary hypothesis, concerning associated of weight reduction with statistically significant diet, genotype pattern interaction and insulin secretion interactions. The authors extensively performed the research while referring to the hypothesis, as evident in measurement of participants’ genotype patter and insulin secretion. Linear, generalized, mixed-effect models were used by the authors to support the hypothesis. Despite the usage of a hypothesis, the authors detected statistically insignificant interactions between diet, insulin secretion and genotype pattern, resulting in an inability of both the hypothesis to support any of these proposed statements (Gardner et al., 2018).

The authors did not adequately discuss the validity of measurement tools and limited the discussion to only mentioning the validity of the hypothetic analysis relying on the assumption that missing data outcomes during follow up are not related to the missing weight values on observed weight and diet treatment variables – hence, indicative of incomplete measurement validity. Internal validity refers to the shortcomings present in the research design itself due to incorrect methodologies and lack of control over additional variables (Steyerberg & Harrell, 2016). Despite lack of study design discussion by the authors, it seems that a prospective study design was adopted which measured the outcomes of weight loss due to dietary changes. ‘Confounding’ or alterations in the dependent variable due to presence of alterative variables apart from independent variables, are a major threat to a study’s findings and hence, internal validity (Andrade, 2018). The authors excluded considerations of additional confounding metabolic and anthropometric variables such as respiratory exchange ratio, body fat percentage and resting energy expenditures, due to lack of adequate findings in the first 78 participants of the study. There was no control group since both populations received differential dietary guidelines. The authors ensured that the dieticians instructing the dietary changes and staff measuring the weight loss outcomes were blinded to the study. However, there was no blinding of the subjects which could prove to be major threat to internal validity due to possible prevalence of bias among subjects (Patino & Ferreira, 2018). Selection bias is a major threat to internal validity, which the authors strictly avoided by selecting participants based on the inclusion and exclusion criteria (Khadka, Burns, G. L., & Becker, S. P. (2016). However, it is worthwhile to mention that the authors level of generalization was high, since the subjects mainly consisted of individuals belonging to high economic and advanced educational backgrounds resulting in greater accessibility to nutritious foods. Lack of inclusion of low educational or low income group backgrounds, may reduce the validity of the research findings. While usage of a selection criteria consisting of characteristics like body mass index and metabolic conditions indicate relation to the study objectives and scenario, the presence of such bias is not indicative of a real life scenario, since obesity is not limited to merely economically and educationally underprivileged groups. An additional threat to the internal validity of the above study is presence of mortality or differential attrition, where authors consider only those subjects, who have endured or successfully completed the study, and not dropouts (Galaviz et al., 2014). The authors mentioned that only 481 out of a total of 609 participants were able to complete the research, with however no mention of the reasons for dropout or whether the values of dropout subjects were concerned, which may pose as a threat to internal validity due to survivorship bias. The authors mention that first 78 participants of the study yielded incomplete findings, for which confounding variables were not considered – a major threat to the validity of this research. In an attempt to reduce this survivorship bias, authors mentioned the occurrence of 9 serious and 11 adverse incidences in this study, resulting in participant illness – but nevertheless, they were still considered. The study included a large sample size, which may result in interval validity threat due to heterogeneity in findings associated with such large populations. External validity pertains to the extent to which the authors generalized their findings and is highly dependent on the presence of internal validity (Andre & Picho, 2016). The population characteristics are key factors affecting generalization and affect external validity (Palinkas et al., 2015). A key threat to the external validity of this study, is the authors’ generalization of including only high income group, educated individuals (Craike et al., 2017). As an implication of such threats, the validity will be affected and the research outcomes will be indicative of biased and incorrect results with lack of inclusiveness of the differential characteristics of individual subjects (Lupton, 2018).  

Selected Participants

Conclusion

The usage of randomization during selection is a key step towards removal of threats to external and internal validity (Palinkas et al., 2015). The authors utilization of a randomized controlled trial, which is considered to posses high validity and level of scientific evidence in evidence based research, was advantageous in removal of potential threats to validity (Andre & Picho, 2016).  The authors should have used randomization during subject selection, through the inclusion of low income groups and individuals of limited educational status, which could have reduced generalization and threats to validity. Authors could have removed threats to validity through the establishment of greater control over the variables – which reduces the effect of confounding variables on the dependent variables (Patino & Ferreira, 2018). The authors should have considered the confounding variables of respiratory exchange ratio, body fat percentage and resting energy expenditure which may have affected the dependent variable of weight loss, rather than the independent variables of diet, genotype pattern and insulin secretion. To achieve management of this threat, the authors could have considered these variables or conducted extensive findings for the 78 participants who presented incomplete values. To conclude, despite the usage of a randomized controlled trial and effective statistical analysis methods, there still remains threats to internal and external validity in this research, which should have been considered by the authors.

References

Andrade, C. (2018). Internal, external, and ecological validity in research design, conduct, and evaluation. Indian journal of psychological medicine, 40(5), 498-499. doi: https://dx.doi.org/10.4103%2FIJPSYM.IJPSYM_334_18.

Craike, M., Hill, B., Gaskin, C. J., & Skouteris, H. (2017). Interventions to improve physical activity during pregnancy: a systematic review on issues of internal and external validity using the RE?AIM framework. BJOG: An International Journal of Obstetrics & Gynaecology, 124(4), 573-583. doi: https://doi.org/10.1111/1471-0528.14276.

Galaviz, K. I., Harden, S. M., Smith, E., Blackman, K. C., Berrey, L. M., Mama, S. K., … & Estabrooks, P. A. (2014). Physical activity promotion in Latin American populations: a systematic review on issues of internal and external validity. International Journal of Behavioral Nutrition and Physical Activity, 11(1), 1-13. doi: https://doi.org/10.1186/1479-5868-11-77.

Gardner, C. D., Trepanowski, J. F., Del Gobbo, L. C., Hauser, M. E., Rigdon, J., Ioannidis, J. P., … & King, A. C. (2018). Effect of low-fat vs low-carbohydrate diet on 12-month weight loss in overweight adults and the association with genotype pattern or insulin secretion: the DIETFITS randomized clinical trial. Jama, 319(7), 667-679. doi: 10.1001/jama.2018.0245.

Khadka, G., Burns, G. L., & Becker, S. P. (2016). Internal and external validity of sluggish cognitive tempo and ADHD inattention dimensions with teacher ratings of Nepali children. Journal of psychopathology and behavioral assessment, 38(3), 433-442. doi: https://doi.org/10.1007/s10862-015-9534-6.

Lupton, D. L. (2018). The External Validity of College Student Subject Pools in Experimental Research: A Cross-Sample Comparison of Treatment Effect Heterogeneity. Political Analysis, 1-8. doi: https://doi.org/10.1017/pan.2018.42.

Palinkas, L. A., Horwitz, S. M., Green, C. A., Wisdom, J. P., Duan, N., & Hoagwood, K. (2015). Purposeful sampling for qualitative data collection and analysis in mixed method implementation research. Administration and Policy in Mental Health and Mental Health Services Research, 42(5), 533-544. doi: https://doi.org/10.1007/s10488-013-0528-y.

Patino, C. M., & Ferreira, J. C. (2018). Internal and external validity: can you apply research study results to your patients?. J Bras Pneumol, 44(3), 183-183. doi: https://dx.doi.org/10.1590/S1806-37562018000000164.

Steyerberg, E. W., & Harrell, F. E. (2016). Prediction models need appropriate internal, internal–external, and external validation. Journal of clinical epidemiology, 69, 245-247. doi: https://doi.org/10.1016/j.jclinepi.2015.04.005.

Torre, D. M., & Picho, K. (2016). Threats to internal and external validity in health professions education research. Academic Medicine, 91(12), e21. doi: 10.1097/ACM.0000000000001446.