This randomized, parallel, sham-controlled trial was conducted at Guang’anmen Hospital (GAMH), China Academy of Chinese Medical Sciences, in Beijing, China. The study course per patient was 24 weeks; baseline measurements were taken at week 0, then followed by 8 weeks of treatment with 20 sessions of acupuncture and 16 weeks of follow-up assessments.

Patients were included if they were diagnosed as having primary SS according to the American–European Consensus Group (11) classification criteria, had a disease course of < 10 years, and were > 18 years old. Moreover, they either had to have not taken any medicines for pSS, or had to have stopped them for ≥ 4 weeks before recruitment, or had to be taking the same kinds and doses of non-steroidal anti-inflammatory drugs, oral glucocorticoids, cholinergic agonists, artificial eye drops, and herbal decoctions for ≥ 4 weeks before recruitment, or had to be taking the same kinds and doses of immunosuppressants (e.g., Tripterygium wilfordii, methotrexate, cyclophosphamide, cyclosporin, azathioprine, and hydroxychloroquine sulfate) for > 6 months before recruitment.

Participants who met one of the following conditions were excluded: had secondary Sjögren’s syndrome; had a serious systemic disorder (e.g., lymphoma; central nervous system, renal, or pulmonary involvement; myositis; or vasculitis) or severe renal or liver failure; had a history of acupuncture treatments within last 20 days; had an acupuncture contraindication (e.g., allergy to metals, skin lesions on relative acupoints, etc.); had participated in any other clinical trials within the last 30 days before recruitment; were planning to be pregnant or were at risk of becoming pregnant (i.e., no contraception use); had a physical or psychological problem that might confound the trial results, interfere with other subjects’ participation, or might make it risky to follow the investigators’ requirements; and were known for or showed persistent drug or alcohol abuse.

Participants were allocated to the acupuncture or sham acupuncture (SA) group with equal chance, based on randomization assignments using a sequence of numbers generated using the Statistical Analysis System version 9.2 (SAS Institute, Cary, NC, United States) by the Institute of Basic Research in Clinical Medicine, China Academy of Chinese Medical Sciences. These assignments were sent to a study member, who placed them into sealed, opaque envelopes with dates and signature labels placed over the seals. This study member subsequently did not take part in any patient interviews, acupuncture sessions, data collection, or statistical analysis. Except for the acupuncturist, all relevant parties were blinded.

During weeks 0–8, 120 patients were randomly treated with acupuncture or SA 3 times a week for the first 4 weeks, then 2 times a week for the next 4 weeks. The needles of the 2 study arms were indistinguishable in appearance and feelings brought to patients. The sham needles were 0.30 mm wide and 25 mm long (12) and consisted of an adhesive pad with a blunt tip. The adhesive pad was made of a sterile cylindrical polyethylene foam (10 mm in diameter and 5 mm in height) with adhesive tape at the bottom. In all, the only difference between the normal and sham needles was that the tips of the sham needles did not taper (13). In the acupuncture group, Hwato acupuncture needles (40 × 0.25 mm; Suzhou Medical Appliance Factory, Suzhou, China) were used. The acupoints used in the acupuncture/SA groups were as follows: bilateral Wai guan (SJ5), bilateral Zhaohai (KI6), Cheng-jiang (RN24), Lianquan (RN23), bilateral Tai yang (EX-HN5), bilateral Cuanzhu (BL2), bilateral Sizhukong (SJ23), and bilateral Jiache (ST6). Equal manipulation via twirling, lifting, and thrusting motions of all needles was performed in the acupuncture group but not in the sham group. Needles in both arms were kept in place for 30 min each session, with the Deqi sensation present only in the acupuncture group (14). The same qualified and registered clinical acupuncturist was trained to maintain consistent performance in both the acupuncture and SA groups by following the standard operating procedure formula. After the 8-week intervention, patients were followed for another 16 weeks to evaluate the long-term effectiveness of treatment.

The primary outcome was the proportion of patients with a ≥ 30% improvement in ≥ 2 of 3 numeric analog scale (NAS) scores for dryness, pain, and fatigue after intervention at week 8 compared to baseline. NAS scores (0–10 points, with 0 points for the best, 10 points for the worst) were requested 3 times a week in week 1–4, and 2 times a week in week 5–8, in total of 20 NAS scores in 8-week intervention, and in week 12, 16, 20, and 24 for follow-up with different measurements. The mean scores of NAS from weeks 2 to 8 were set as the primary endpoint, while those of week 0 were considered the baseline NAS scores. The methodology and evidence of setting the requirement of a ≥ 30% improvement in ≥ 2 of 3 indicators is closely related to clinical efficacy and are consistent or similar to the main results of multiple previous clinical trials (5, 8, 15).

The secondary outcomes included NAS scores (mean ± standard deviation [M ± SD]) for dryness, pain, and fatigue as well as the European League Against Rheumatism (EULAR) Sjogren’s syndrome Patient-reported Index (ESSPRI); the EULAR Sjögren’s Syndrome Disease Activity Index (ESSDAI) (16); the Medical Outcome Study Short Form 36 (SF-36) score; the Hospital Anxiety and Depression Scale (HADS) score (17); the Schirmer test score; unstimulated saliva flow; the erythrocyte sedimentation rate (ESR); the C-reactive protein (CRP) level; serum immunoglobulin (Ig) G, IgA, and IgM levels; and an ultrasound examination of the salivary gland structures.

All salivary gland ultrasonography examinations were performed independently by the same examiner with professional training and sufficient operating experience. Homogeneity, the presence of hypoechogenic areas and hyperechogenic reflections, and the clearness of salivary gland borders were scored according to the scoring system proposed by De Vita (18, 19). The scores of these 4 parameters for the 4 salivary glands were then summed up, ranging from 0 to 48 points for each. Scores calculated at weeks 0 and 8, respectively, were compared.

The sample size was pre-decided by the hypothesized primary outcome. According to an existing literature review (5), the proportion of patients who reported a ≥ 30% improvement in NAS scores of ≥ 2 of the 3 items with placebo treatment was 17.9%. We used this proportion for our placebo group because no data were available for SA at the time. A sample size of 52 patients per group would be sufficient to detect a 2% null difference in proportions, with a 2-sided 5% level of significance and a power of 80%. The number of subjects was increased to 60 participants in each group (for a total of 120 participants) when a dropout rate of 15% was taken into consideration.

Statisticians and investigators were blinded to the patients’ allocation. Statistical analysis was performed using Stata version 15.1 (StataCorp LLC, College Station, TX, United States) and GraphPad Prism version 8.4 (GraphPad Software, San Diego, CA, United States). We used the first ESSPRI and NAS scores as the baseline scores. Thus, participants without a single session of acupuncture were excluded from statistical analysis. Missing data for the primary outcome were extrapolated using the last-observation-carried-forward (LOCF) approach.

Continuous data are presented using M ± SD values if they were normally distributed; otherwise, they are presented as median ± interquartile range values. Categorical data are presented as frequencies and percentages. In comparisons between the 2 groups, the chi-squared test or Fisher’s exact test was used for primary outcomes. For secondary outcomes, a t-test was used if the data were normally distributed; otherwise, the Wilcoxon rank-sum test was applied. An analysis of covariance was used for the baseline adjustment analysis of continuous data. For repeated outcome measurements, a general linear model was used for analyzing continuous data.

The threshold of statistical significance of the 2-tailed P-value was 0.05. For repeated outcome measurements, the P-values were corrected using the Huynh–Feldt adjustment method.

From March 2016 to September 2018, 120 patients with pSS treated at GAMH were recruited into this trial (ClinicalTrials.gov identifier: NCT02691377). The last follow-up assessment of the last patient was completed on March 11, 2019.

All 120 patients were equally (60 patients per group) and randomly assigned to the acupuncture or SA group. The mean age was 56.7 ± 8.2 years in the SA group and 53.8 ± 12.1 years in the acupuncture group, and 88.3% of patients in the SA group and 98.3% of patients in the acupuncture group were female. The disease course of pSS was 35.3 ± 29.4 months in the SA group and 28.9 ± 23.4 months in the acupuncture group, respectively. There were no significant differences in any characteristics between the 2 groups at baseline. Fourteen participants (11.67%) at week 8 and 30 patients (33.3%) at week 24 were lost to follow-up; these losses to follow-up and their reasons are further described in Figure 1. Study participants’ demographic characteristics and concomitant treatment are summarized in Table 1 and their ESSDAI domains are discussed in Table 2.

At week 8, the proportions of patients meeting the primary endpoint (≥ 30% reduction in ≥ 2 of the 3 NAS scores for dryness, pain, and fatigue) were 28.33% (17/60) in the acupuncture group and 31.66% (19/60) in the SA group, without a statistically significant difference (P = 0.705 after LOCF imputation) (Table 3).

In addition, we also compared the proportions of patients who experienced a ≥ 30% reduction in 2 of the 3 NAS scores in the 2 groups in other weeks during follow-up (Table 3), including the actual values of the odds ratios. However, no significant difference between the groups was obtained at any week.

No difference in NAS scores for dryness, fatigue, or pain was observed between the 2 groups during any week. However, there was a general decreasing trend in NAS scores observed in both groups during the intervention and follow-up periods (Table 4 and Figure 2).

In a post hoc analysis, the minimally clinically important improvements (MCIIs) in NAS scores for dryness, pain, and fatigue were calculated as −1, −2, and −3 point(s) on 10-point scales, respectively. There was no significant difference between the 2 groups in the proportion of patients with MCII scores at week 8 in each of the 3 scales or in ≥ 2 of 3 of the scales, nor was there a significant difference in the proportion of patients with an improvement of 1, 2, or 3 point(s) in the NAS scores for pain, fatigue, or dryness or in ≥ 2 of 3 of the scores (Table 5).

In week 16, the IgG concentration (M ± SD) of the SA group showed a significant difference (P = 0.0490) compared to that of the acupuncture group (17.65 ± 7.20 g/L vs. 16.97 ± 5.46 g/L) (Table 4). No other difference in ESSPRI, ESSDAI, SF-36, the HADS anxiety subscore, the HADS depression subscore, Schirmer test results, unstimulated saliva flow, ESR, the CRP level, or the IgA or IgM concentration between the 2 groups was obtained in any other week. Compared to week 0, there was a downward trend in unstimulated salivary flow in both groups, more pronounced in the acupuncture group, but no significant difference was found between the two groups (Figure 3). In exploring the differences between before and after the intervention period in the 2 groups, we observed significant differences in ESSPRI between weeks 0 and 8, weeks 0 and 12, weeks 0 and 16, and weeks 0 and 24 between the groups and noted significant differences in unstimulated saliva flow between weeks 0 and 8, weeks 0 and 16, and weeks 0 and 24 between the groups (Table 6).

We quantified and calculated the ultrasonography improvement ratio of the 2 groups before and after the 8-week intervention in 4 different dimensions of ultrasonography (18, 19). Both groups acquired improvements in salivary gland ultrasound scores, with a 18.7% improvement in the SA group and a 35.2% improvement in the acupuncture group noted in terms of homogeneity, respectively, and these were significantly different results (P = 0.0334). No significant difference was obtained in the other 3 dimensions (Table 7).

To assess blinding, in week 8, we asked patients if they felt a “deep penetration,” “shallow penetration,” or “no penetration.” Choosing either former two options indicated successful blinding. None of the participants answered “no penetration,” which suggested that the blinding was successful. 78.1% of patients in acupuncture group reported that they had deep penetration, and 60.0% in sham acupuncture group (P = 0.122). Additionally, 78.0% of patients were willing to recommend acupuncture treatment to others, showing a good satisfaction.

During the first 8 weeks, there was 1 serious adverse event in the placebo group and 2 in the acupuncture group. In the last 16 weeks, there was 1 serious adverse event in the placebo group and none in the acupuncture group. More details of the adverse events are described in Table 8.