HTN: Methods (2022-24)

HTN: Methods (2022-24)

Methods for Systematic Review Examining the Effect of Dietitian Interventions on Adults with
Pre-hypertension or Hypertension

This systematic review and meta-analysis is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist and follows the Cochrane Collaboration’s Grading of Recommendations, Assessment, and Evaluation (GRADE) methodology1,2. Research questions, eligibility criteria, search strategy, and methods of screening, data extraction and data analysis were determined a priori and registered on PROSPERO (CRD42022351693)3

Search Strategy

A systematic search of the electronic databases MEDLINE, CINAHL, and Cochrane CENTRAL was conducted by an information specialist for articles published from 1985 to August 2022. The search was limited to studies from peer-reviewed journals with titles and abstracts available in English. Reference lists of relevant systematic reviews and primary articles were hand searched to identify potentially included studies that may not have appeared in the database searches.

Eligibility Criteria and Study Selection

The eligibility criteria were developed using the Population, Intervention, Comparators, and Outcome (PICO) criteria. Included studies must have implemented an MNT intervention (nutrition intervention provided by a dietitian) compared to standard care group or no intervention, herein referred to as ‘control group’, among free-living adults (≥18 years old) with pre-hypertension or hypertension. The dietitian could provide MNT alone or as a part of a multi-disciplinary intervention, and interventions were required to be four weeks or longer in duration with at least two dietitian contacts. Primary outcomes of interest included SBP and DBP, hypertension incidence, and changes in anti-hypertension medication usage. Secondary outcomes included CVD events (stroke, myocardial infarction), calculated 10-year CVD risk score, cardiovascular-related hospitalizations, intima-media thickness/arterial stiffness/endothelial function, anthropometrics (body weight, waist circumference, body mass index (BMI)), mortality (all-cause or CVD-mortality), quality of life, cost-effectiveness, and adverse events. Included articles were limited to randomized controlled trials (RCTs), as this is the most rigorous study design to test intervention efficacy. 

Abstracts were independently screened by two reviewers using the Rayyan database screening software4. Reviewers discussed any discrepancies before establishing consensus. Full texts of potentially eligible articles were then retrieved and independently evaluated against the eligibility criteria by two reviewers. Similarly, any discrepancies were discussed between the two reviewers until consensus was reached or resolved by content experts. If additional information was needed to determine a study’s eligibility, the study’s corresponding author was contacted. If data reported in the included studies was too sparse or heterogeneous to be included in meta-analysis, results were reported by narrative synthesis only.

Data Extraction

Data was extracted from eligible studies onto a standardized form by an analyst and accuracy was confirmed by a second reviewer. Extracted data included participant characteristics (e.g., age, sex, BP thresholds for pre-hypertension or hypertension diagnosis, anti-hypertension medication use), MNT intervention details (e.g., duration, number and frequency of contacts), comparator description, outcomes of interest reported, intervention setting, and funding information. Quantitative data extracted for meta-analysis included sample size, mean difference in outcome, and variance for each group for the continuous variables, and the number of events and sample size for each group for the categorical variables. 

Risk of Bias Assessment

Risk of bias for each study was assessed independently by two reviewers using the Cochrane Risk of Bias tool, Version 2 (RoB 2)5. This tool assesses five domains of bias arising from the randomization process, deviations from the intended intervention, missing outcome data, measurement of the outcome, and selection of reported results. Discussion was held among reviewers and any discrepancies were considered before establishing consensus. Risk of bias was reported using the Robvis data visualization tool6.

Data Synthesis

Study and intervention characteristics for all included studies were summarized in tables. A meta-analysis was conducted when quantitative data for an outcome was available from two or more studies that reported data similar enough to be pooled. If a study reported outcomes at multiple time points, data nearest to the intervention endpoint were used for the meta-analysis. Continuous outcomes were reported as mean difference when measured using the same tool and/or unit, or standardized mean difference for continuous outcomes measured using different tools and/or units, and 95% confidence interval (95% CI). Categorical outcomes were reported as relative risk and 95% CI. For studies that had multiple MNT intervention groups (e.g., two or more MNT intervention groups in which the diet prescribed differed), each MNT intervention was independently compared to the control group. Meta-analyses were performed using a random-effects model in OpenMeta [Analyst] and forest plots were generated7. Heterogeneity was determined by the I2 statistic and further explored by subgroup analyses (e.g., year of publication, intervention duration, anti-hypertension medication use, overweight/obesity status, if the dietitian was the sole healthcare provider or part of a multidisciplinary intervention team, if care was provided individually or in a group). Sensitivity analysis was conducted by comparing effects according to study quality. When a meta-analysis included ≥10 studies for an outcome, publication bias was assessed using the Egger’s statistic and by generating funnel plots8.

Certainty Assessment

Certainty of evidence was determined using the GRADE method and was dependent upon study design, sample size, collective risk of bias, inconsistency in findings, indirectness, imprecision, and other factors9. Certainty of evidence was graded as “high”, “moderate”, “low”, or “very low” and presented in the summary of findings table10


  1. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Annals Intern Med. 2009;151(4):W-65-W-94.
  2. Moberg J, Oxman AD, Rosenbaum S, et al. The GRADE Evidence to Decision (EtD) framework for health system and public health decisions. Health research policy and systems. 2018;16(1):1-15.
  3. Katelyn Senkus, Josephine Dudzik, Mary Rozga, Lisa Maloney, Deepa Handu. Effectiveness of medical nutrition therapy provided by a registered dietitian nutritionist in adults with pre-hypertension and hypertension: A systematic review. PROSPERO 2022 CRD42022351693 Available from:
  4. Ouzzani, Hammady, Fedorowicz, Elmagarmid. Rayyan – a web and mobile app for systematic reviews. Syst Rev. 2016; 5(1): 210.
  5. Sterne JA, Savovic J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366.
  6. McGuinness LA, Higgins JPT. Risk-of-bias VISualization (robvis): an R package and Shiny web app for visualizing risk-of-bias assessments. Res Synth Methods. 2021; 12 (1): 55-61.
  7. Wallace BC, Dahabreh IJ, Trikalinos TA, Lau J, Trow P, Schmid CH. Closing the gap between methodologists and end-users: R as a computational back-end. J Statis Software. 2012;49:1-15.
  8. RStudio Team (2020). RStudio: Integrated Development for R. RStudio, PBC, Boston, MA
  9. RStudio Team (2020). RStudio: Integrated Development for R. RStudio, PBC, Boston, MA
  10. G.D.T. GRADEpro. GRADEpro Guideline Development Tool [Software] McMaster University and Evidence Prime (2022) Available from: