We recruited hypertensive patients from August, 2019 from two community centers in Dalian City, Liaoning Province, China. Hypertensive status was defined as systolic BP (SBP) $\ge$140 mmHg and/or diastolic BP (DBP) $\ge$90 mmHg or taking antihypertensive agents [7]. Inclusion criteria were: (1) primary hypertensives with 140 $\le$ SBP $\le$ 180 mmHg and/or 90 $\le$ DBP $\le$ 110 mmHg measured by professional doctors; (2) aged $\ge$50 and $\le$70 years; (3) at least had two meals at home per day; (4) had no severe renal insufficiency; (5) baseline serum potassium concentration lower than 5.0 mmol/L.

This 12-month study utilized a computerized randomization program to divide participants into two groups in a 1:1 ratio into the normal salt (NS) group or the LS group. The NS was comprised of 100% sodium chloride, while the LS substitution contained 43% sodium chloride, 32% potassium chloride and 25% other ingredients.

Fig. 1 presents the flow chart of the study. The baseline investigation included interviews of all participants conducted by professional doctors to gather demographic characteristics from all participants.

Fig. 1.

Flow chart of the study. NS, normal salt; LS, low-sodium; ABPM, ambulatory blood pressure measurement; PRC, plasma renin concentration; Ald, aldosterone; hs-CRP, high-sensitivity C-reactive protein; MALB, microalbuminuria; UCRE, urine creatinine.

Baseline and endpoint ABPM were taken in the right arm using a validated electronic upper-arm cuff device (ABPM50, Qinhuangdao, China). Daytime measurements were taken at 30-min intervals from 6:00 AM to 10:00 PM; nighttime measurements were taken at 60-min intervals from 10:00 PM to 6:00 AM. Patients were informed to continue their daily activities but avoid vigorous activity. At least 20 valid daytime and 7 nighttime BP readings were required for the measurements to be considered successful. If not, participants were scheduled to take the exam again [7], and they were excluded if there were insufficient valid readings.

Fasting blood samples were collected to measure plasma renin concentration (PRC), aldosterone (Ald), and hs-CRP along with first morning urine samples for measuring microalbuminuria (MALB), urine creatinine (UCRE), and spot urinary sodium and potassium, the latter to ensure compliance with the LS salt substitution regimen [14]. These samples were obtained from each participant at baseline and at the study’s endpoint. Subsequently, all blood and urine specimens were sent to the clinical laboratory of Dalian Medical University (DMU) for analysis.

We used the following formula to calculate the sample size:

$$\mathrm{N}=(1+1/\kappa ){\left(\sigma \frac{\mathrm{z1}-\frac{\alpha }{2}+\mathrm{z1}-\beta }{\mu \mathrm{A}-\mu \mathrm{B}}\right)}^2$$

the value of $\alpha$ was 0.05 and $\beta$ was 0.1. The minimal sample size was 160 in both groups. This number was increased by 10% to compensate for potential losses due to follow-up or other reasons. Hence, the final sample size was 352 patients.

We conducted statistical analyses using SPSS 25.0 (IBM Corp., Armonk, NY, USA). We reported data as mean $\pm$ SD, and t-tests were used to compare continuous variables with a normal distribution. For variables with an abnormal distribution, median and inter-quartile ranges were used, and the Mann–Whitney U test was applied for comparisons. Frequencies were used to describe categorical variables, and the Chi-squared test was employed for comparisons. A p value of 0.05 was considered statistically significant.

Initially, 352 participants were enrolled in the study. Of these, 30 (8.5%) dropped out, leaving 322 who completed all follow-up visits, and 301 who completed the ABPM. The average participant age was 62.17 $\pm$ 4.69 years in the NS group and 62.96 $\pm$ 4.51 years in the LS group. There were 72 (45.0%) male subjects in the NS group and 58 (35.8%) in the LS group. No significant differences were recorded in baseline age, sex, body mass index (BMI), and antihypertensive medication (all p $>$ 0.05) (Table 1).

At baseline, there were no significant differences in ABPM between the two groups for any period (all p $>$ 0.05). However, at the study endpoint, there were noteworthy changes in blood pressure levels. The 24-hour diastolic blood pressure (24hDBP) was significantly lower in the LS group when compared to the NS group (p = 0.008). Additionally, there was a non-significant trend towards a decrease in the 24-hr systolic BP (24hSBP) compared to the NS group.

The daytime DBP (dDBP) was significantly lower in the LS group (p = 0.015). The daytime SBP (dSBP) also exhibited a non-significant trend towards a decrease in the LS group. Despite the presence of a trend towards a decrease, neither the nighttime SBP (nSBP) or nighttime DBP (nDBP) reached statistical significance (Fig. 2).

Fig. 2.

Comparison of ambulatory BP measurement (ABPM) between NS and LS group at baseline (2a) and the endpoint (2b). * p 0.05. NS, normal salt; LS, low-sodium; SBP, systolic blood pressure; DBP, diastolic blood pressure; d, daytime; BP, blood pressure; n, nighttime.

Examining the difference from baseline yielded notable results as well. We found significant differences in both $\mathrm{\Delta }$24hSBP (p = 0.004) and $\mathrm{\Delta }$24hDBP (p = 0.007) in the LS cohort when compared to the NS cohort. Additionally, $\mathrm{\Delta }$dSBP (p = 0.012), $\mathrm{\Delta }$dDBP (p = 0.012), and $\mathrm{\Delta }$nSBP (p = 0.007) differed significantly as well. In contrast, there were no differences between groups when assessing $\mathrm{\Delta }$nDBP (Table 2).

Participants were divided into tertiles based on their PRC. In the first tertile of the PRC subgroup (PRC: $\le$5.896 uIU/mL), there was a notable decrease in 24hSBP and 24hDBP by –7.5 mmHg and –5.2 mmHg, respectively. In the second tertile (PRC: 5.896–17.040 uIU/mL) the changes were minimal at 0.3/–1 mmHg for 24hSBP and 24hDBP. For the third tertile (PRC: $>$17.040 uIU/mL) the changes were –1.2/0.4 mmHg. Daytime values for dSBP and dDBP also varied across tertiles. In the first, second, and third tertiles, the changes were –5.7/–6 mmHg, –1.5/0.3 mmHg, and –0.7/0.2 mmHg, respectively. Nighttime systolic and diastolic blood pressure (nSBP and nDBP) followed a similar pattern, with changes of –6/–3.5 mmHg in the first tertile, 1.4/0.1 mmHg in the second, and 0.1/1.4 mmHg in the third. Comparatively, blood pressure changes were more pronounced in the low PRC group than in the high PRC group. This difference was statistically significant for all measurements: 24hSBP (p = 0.001), 24hDBP (p = 0.003), dSBP (p = 0.011), dDBP (p = 0.006), nSBP (p = 0.003), and nDBP (p = 0.006) as shown in Fig. 3a.

Fig. 3.

Comparison of changes in ABPM for different subgroups according to baseline PRC (3a), aldosterone concentration (3b) and ARR (3c). $\mathrm{\Delta }$, net difference; * p 0.05; ** p 0.001; Ald, aldosterone; ARR, aldosterone/renin ratio; PRC, plasma renin concentration; SBP, systolic blood pressure; DBP, diastolic blood pressure; n, nighttime; d, daytime.

In the first tertile of aldosterone subgroups (Ald: $\le$84.000 pg/mL), changes in 24-hr BP were modest, at –1.6/–1.1 mmHg for systolic and diastolic pressures, respectively. The second tertile (Ald: 84.000–130.000 pg/mL) showed slightly greater changes at –3.3/–2.3 mmHg. However, in the third tertile (Ald $>$130.000 pg/mL) the changes were –1.2/0.7 mmHg, indicating lesser change in systolic and an increase in diastolic pressures. Changes in daytime BP followed a similar pattern, with the first, second, and third tertiles showing changes of –1.9/0.4 mmHg, –3.1/–2 mmHg, and –3.2/–1.4 mmHg, respectively. For nighttime BP, the changes were –1.3/–1 mmHg in the first tertile, slightly different at –0.4/–1.6 mmHg in the second, and 0.7/1.2 mmHg in the third, indicating an increase in both systolic and diastolic pressures for the highest aldosterone subgroup. Despite these variations, no significant differences were observed between the three subgroups for any of the blood pressure changes (p = 0.643 for 24hSBP, p = 0.184 for 24hDBP, p = 0.622 for dSBP, p = 0.127 for dDBP, p = 0.401 for nSBP, and p = 0.165 for nDBP), as illustrated in Fig. 3b.

In the first tertile of the ARR subgroup (ARR: $\le$5.420), 24-hr BP changes were minimal, at –0.8/–0.3 mmHg for systolic and diastolic pressures, respectively. In the second tertile (ARR: 5.420–16.550) the changes were even more modest at –0.1/0.3 mmHg. However, the third tertile (ARR: $>$16.550) exhibited substantial changes, with reductions of –6.6/–5.2 mmHg. Daytime dBP changes also varied across tertiles. The first, second, and third tertiles showed changes of –0.5/0.6 mmHg, –1.6/0.4 mmHg, and –5.2/–5.4 mmHg, respectively. Nighttime nBP changes followed a similar trend, with the first and second tertiles showing minimal changes of –0.2/0.3 mmHg and 3.1/1.2 mmHg, respectively, and the third tertile showing a significant reduction of –5/–3.8 mmHg. Notably, BP changes in the high ARR group were significantly greater than those in the low ARR group. This was evident across all measurements (p = 0.001 for change in 24hSBP, p 0.001 for change in 24hDBP, p = 0.012 for change in dSBP, p = 0.001 for change in dDBP, p = 0.008 for change in nSBP, and p = 0.006 for change in nDBP) (Fig. 3c).

At baseline, there were no significant differences were found between the two groups in terms of MALB, UCRE, and UACR. By the study endpoint, notable differences emerged. Specifically, the UACR was significantly lower participants assigned to the LS cohort, with median values of 12.48 µg/mg (interquartile range [IQR]: 5.12, 26.46) vs. 19.63 µg/mg (IQR: 7.65, 36.94) and a p value of 0.026 indicating statistical significance. Meanwhile, no significant differences were observed in MALB and UCRE between the two groups at the endpoint (Table 3).

There were no differences in the changes of MALB and UCRE in the two groups. In contrast, we did observe a significant difference in the UACR change between groups (p = 0.004) (Table 4).

At baseline, we found no differences in hs-CRP between the NS and LS groups (1.16 mg/L [IQR: 0.56, 2.61] vs. 1.26 mg/L [IQR: 0.61, 2.2], p = 0.999). In contrast, a reduction in hs-CRP was observed at the study endpoint (0.73 mg/L [IQR: 0.40, 1.38] vs. 0.87 mg/L [IQR: 0.52, 1.79], p = 0.045 (Fig. 4a). This effect was also found when assessing hs-CRP change (–0.24 mg/L [IQR: –0.96, 0.13] vs. –0.06 mg/L [IQR: –0.93, 0.36], p = 0.048) (Fig. 4b).

Fig. 4.

Hs-CRP at baseline and the endpoint (4a) and comparison of change of hs-CRP (endpoint to baseline) between NS and LS group (4b). * p 0.05. NS, normal salt; LS, low-sodium; $\mathrm{\Delta }$, net difference; hs-CRP, high-sensitivity C-reactive protein.