Discussion
Our post hoc statistical model that allows for the variation in effect by the time of intervention, which may fit better the nature of the intervention, showed a favorable effect of the progressive reduction of salt supply on both SBP and DBP at the end of 2 years of intervention, although this was not seen in the analysis per protocol. While 24-h urinary sodium excretion from a small subset of study participants was not reduced in the intervention arm, the effects on self-reported food tasting and self-addition of out-of-study table salt in a larger subset, as well as the results from the analysis excluding those who reported self-addition of out-of-study table salt, suggest that sodium intake in the PR group may have been reduced and may have contributed to the observed blood pressure differences. However, these results differed from the pre-specified model that respected the randomization but assumed an intervention effect that did not vary over time [
11]. It is therefore important to outline some aspects of the incremental sodium reduction implementation in DECIDE-Salt to help us understand the intervention.
The premise of the strategy of gradual incremental reduction of dietary salt is that this policy could be introduced unobtrusively for a population, or a group. The sodium content of foods could then be reduced in a progressive step-wise manner, allowing adaptation of consumers’ taste before each step [
16,
17]. This was the intent of the salt reduction intervention in this trial.
However, in the particular circumstances of DECIDE-Salt, from early in the trial self-reports of food tasting bland as well as action of some participants to add out-of-study table salt, both occurred more frequently in the intervention than in the control arm—supporting two important conclusions: First, salt reduction appeared to be achieved by the intervention but, second, the intent to introduce salt reduction unobtrusively in the whole diet was not fully successful. The strategy applied in prior studies was generally limited to a single food such as bread or only a few selected foods, perhaps making the reduction of salt content less likely to be recognized [
8,
9,
17]. To our knowledge, the DECIDE-Salt trial was the first to try the stepwise approach to lower salt use in the whole diet.
Similar, though less frequent, self-reports of blandness of the food were received from participants in the control arm. It is possible that some of the control facilities may have adopted a salt reduction policy on their own, in disregard of the protocol. Our findings from the case studies, which integrated the information on self-reported blandness of food, urinary sodium, and blood pressure, indicated that salt reduction may have taken place in some control facilities. The high proportion of participants reporting food blandness at baseline in both intervention and control groups also supports this idea. Although this did not affect the balance of intervention groups because the randomization was done after baseline, it may have reduced the potential impact of the salt reduction intervention, leaving less room for the intervention to take effect. For future studies, to prevent such “spontaneous intervention” before the actual intervention, the kick-off meeting and training of the facilities’ managers should emphasize that they should strictly follow the study protocol to implement the study interventions assigned to them.
A factor limiting the evaluation of the salt reduction intervention was the health status of the study participants. Co-morbidity frequently limited the completeness of follow-up. The reduced effective sample size and observed numbers of detectable events may have limited the statistical power of the study to identify potential differences in outcomes, e.g., 24-h urinary sodium, for which only 639 participants had data at 24 months for analysis.
Despite these difficulties, the study achieved over 75% of participants with follow-up blood pressure measured for the assessment of effects on primary outcome, and the characteristics of those participating were similar between the study arms, as noted above.
The difficulty we had in reducing sodium intake and blood pressure through the progressive reduction of salt supply demonstrated that it is not easily achieved and may help to explain why the WHO global target for population sodium reduction by 30% by 2025 would not be achieved [
18] if no effective strategies other than mass health education is developed and rolled out globally.
Long-term reduction of salt intake requires effort to achieve the change in dietary behavior of the consumers unless the change cannot be detected, the change is within the tolerable range, or the consumers are prepared and ready for the change. In the DECIDE-Salt trial, only a simple health education program was given, and the participants may have been poorly prepared for the intervention. In such circumstances, the detection of salt reduction in foods by the consumers through the accompanying change in food taste might lead to their opposition to the intervention [
19,
20]. We would suggest to add in a health education program before initiating the salt supply reduction program so that the facilities’ residents are well prepared for and may more readily accept the intervention, and briefly but continuously reinforce the messaging during the implementation of the salt supply reduction program to help to maintain enthusiasm for the intervention.
Further insights into the strategy of unobtrusive salt reduction are important for the design and implementation of future studies. (1) Where cooperation of an institutional gatekeeper such as kitchen management is required, thorough orientation to the study is needed for those involved in both arms of the trial—its rationale, potential benefits, and need to adhere strictly to the protocol to permit evaluation. (2) Persistence of study oversight and monitoring of fidelity to the intervention design are necessary over the full course of the trial. (3) To permit ongoing assessment of changes in salt intake, a subsample of willing participants should provide 24-h urine collections at each step for analysis of sodium excretion. (4) An interesting question for further study is whether seasonal variation in climate is associated with customary changes in salt intake or salt taste [
21]. Due to Chinese culture and the variability in the availability of fresh foods between seasons, people in northern China like to have low salt, low-fat dishes with more fresh vegetables and fruits in warm seasons and warm and meaty dishes that require more salt added to satisfy the same taste in cold seasons; for trial settings with collective catering, the answer may be to suggest special measures over the course of study, e.g., different diets for summer and winter seasons. Also important is the assurance of comparability between intervention groups in the timing of evaluations with respect to the season. (5) Composition of a candidate study population should be considered with respect to age, sex, comorbidities, and other factors in order to optimize compliance [
22], outcome assessment, and generalization to other populations. (6) Consideration should be given to extending post-intervention follow-up to assess the sustainability of the intervention and its benefits long-term. An alternative, and potentially complementary strategy, is the provision of sodium-reduced, potassium-enhanced table salt for use in cooking and at table, which has been shown in the SSaSS trial [
23] in China to reduce both blood pressure and cardiovascular events.
Acknowledgements
The authors thank all facility residents for their participation and cooperation. We also like to thank all investigators, study team members, facility managers and staff, and administrative agencies for their proactive participation in the study. We are also grateful to Prof Kiang Liu at the Northwestern University School of Medicine for his advice on study design and statistical analysis plan development.
Group information
Management committee
Yangfeng Wu (Chair), Peking University Clinical Research Institute and School of Public Health, Beijing, China.
Xiangxian Feng, Department of Preventive Medicine, Changzhi Medical College, Changzhi, Shanxi, China.
Ruijuan Zhang, School of Public Health, Xi’an Jiaotong University, Xi’an, Shaanxi, China.
Hongxia Wang, Hohhot Center for Disease Control and Prevention, Hohhot, Inner Mongolia, China.
Qianku Qiao, Yangcheng Ophthalmology Hospital, Shanxi, China.
DECIDE-salt study advisory committee
Junshi Chen (Chair), China National Food Safety Risk Assessment Center, Beijing, China.
Bruce Neal, The George Institute for Global Health, University of New South Wales, Sydney, Australia.
Darwin Labarthe, Northwestern University Feinberg School of Medicine, Chicago, United States.
Minghui Zhao, Renal Division, Department of Medicine, Peking University First Hospital, Beijing, China.
Wenyi Niu, Department of Social Medicine and Health Education, Peking University School of Public Health, Beijing, China.
Hai Fang, China Center for Health Development Studies, Peking University, Beijing, China.
Runlin Gao, Department of Cardiology, Fuwai Hospital, Peking Union Medical College & Chinese Academy of Medical Sciences, Beijing, China.
Xiaofeng Liang, Chinese Preventive Medicine Association, Beijing, China.
Yong Huo, Department of Cardiology, Peking University First Hospital, Beijing, China.
Changsheng Ma, Department of Cardiology, Beijing Anzhen Hospital, Capital Medical University, Beijing, China.
DECIDE-salt study data monitoring and safety board
Yihong Sun (Chair), Department of Cardiology, China-Japan Friendship Hospital, Beijing, China.
Xun Tang, Department of Epidemiology and Biostatistics, Peking University School of Public Health, Beijing, China.
Wei Zhao, Department of Cardiology, Peking University Third Hospital, Beijing, China.
Zhun Sui, Department of Nephrology, Peking University People’s Hospital, Beijing, China.
Jinwei Wang, Department of Medicine, Peking University First Hospital, Beijing, China.
DECIDE-salt study clinical outcomes adjudication committee
Xingshan Zhao (Chair), Department of Cardiology, Beijing Jishuitan Hospital, Beijing, China.
Xin Du, Department of Cardiology, Beijing Anzhen Hospital, Capital Medical University, Beijing, China.
Weiping Sun, Department of Neurology, Peking University First Hospital, Beijing, China.
DECIDE-salt collaborative study participanting centers and staff information
Peking University Clinical Research Institute (Study coordinating center): Yangfeng Wu (PI), Huijuan Li (Project Manager), Aoming Jin (Study Coordinator), Jiayu Li (Study Manager), Gaoqiang Xie (Data Management Supervisor), Fengzhi Wang (Data Manager), Shulan Zhu, Yifang Yuan, Lin Feng.
Changzhi Medical College: Xiangxian Feng(KI), Jianhui Yuan, Peifen Duan, Yanbo Han, Xiaorui Cui, Decheng Li, Peng Gao.
Yangcheng Ophthalmology Hospital: Qianku Qiao(KI), La’e Cao, Lili Cheng, Lili Zhang, Lidong JI, Wenjie Yu.
Hohhot Center for Disease Control and Prevention: Hongxia Wang(KI), Hui Zhang, Sujuan Wang, Ran Zhuo, Jing Hang, Li Yang, Xiaoqin QI.
Xi’an Jiaotong University: Ruijuan Zhang(KI), Senke Hu, Yanxi Wang, Yuqi Wang, Yang Shen, Huan Chang.