This study was part of a three-arm randomized double-blinded controlled trial (ClinicalTrial.gov: NCT02498795) (24). The study followed the Helsinki Declaration and was approved by the local Ethics Committee (C.P.−C.I. PI15/0017), and all participants consented their enrolling in this study.

Adults aged between 18 and 45 years were recruited from various sports clubs and federations. To be eligible, participants had to meet specific inclusion criteria: (1) experienced anterior knee pain below the patella while engaging in sports for over 3 months; (2) engaged in sports activities at least 3 times a week; and (3) scored below 80 on the Victorian Institute of Sport Assessment-Patellar questionnaire (VISA-p). Exclusion criteria included: (1) knee surgery in the past 6 months; (2) patellar tendon corticosteroid injection in the past 3 months; (3) diagnosed with chronic joint disease; (4) contraindications for needling (e.g., needle phobia, needle material allergy); (5) consumption of anti-inflammatory, analgesic, or antibiotic medications within the past 48 h; and (6) undergoing concurrent physiotherapy treatment. Additionally, ultrasound examination of the knee joint and adjacent musculoskeletal structures was conducted before enrollment in the study. This examination aimed to exclude the presence of joint effusion or signs of inflammation and to identify the presence of degenerative signs, characterized by a hypoechoic area in the body of the tendon. None of the participants in the final selected sample had received prior needling treatment in the tendon.

Participants were divided into three groups according to the intervention received: (1) PNE [no-sham group], (2) sham PNE by using DN [single-sham group], and (3) sham PNE by using sham needles [double-sham group]. The setting and procedure followed with each participant was similar regardless of group, isolating the effects related to galvanic current and needling (Figure 1). All interventions were targeted at the patellar tendon.

Participants underwent 4 intervention sessions over an 8-week period, with each session spaced 2 weeks apart. During each session, a standardized procedure was followed to ensure participant blinding during needle interventions, attempting to overcome biases found in previous studies on needling techniques (25). Participants were positioned supine with their knee flexed at 20°, supported by a pillow. The researcher performing the intervention wore latex gloves and cleansed the area with a 70% propan-2-ol antiseptic solution. A disposable protective cover was applied to a lubricated ultrasound probe (Logic S7 Expert, General Electric Healthcare), which was used for real-time ultrasound guidance. The ultrasound display screen was positioned behind the participants. Each participant held the anode connected to the electrolysis device and received the following instruction: “During the needle intervention, please try to remain still. The treatment may cause some pain or discomfort. If you experience any, please let me know, and I will stop immediately.” After needle removal, the area was gently compressed with cotton wool for 5 s (Figure 1).

At the end of the study, participants were asked via email to guess the type of treatment they received. The options provided were: “No needling treatment,” “Needling treatment,” or “I don’t know.” If participants selected “Needling treatment,” they were further asked to specify whether they received PNE or DN, with the options being “PNE,” “DN,” or “I don’t know.”

Complementary to the real or sham needle intervention, all participants were instructed to perform a unilateral eccentric exercise program of the quadriceps muscle on the affected side, specifically aimed at the patellar tendon. This program consisted of performing 3 sets of 15 repetitions daily on a decline board (26). The correct execution of the exercise as well as the follow-up of the prescribed program was monitored by the research team every two weeks, coinciding with the day the participant received the intervention.

Participants who fulfilled the inclusion criteria and consented to participate in the trial were randomly assigned by a researcher not involved in the study by generating random participant sequences with a 1:1:1 allocation using an opaque envelope, with a block size of 15 participants, using a computer program (Randomizer).¹

Assessments were made by an assessor blinded to group allocation. The primary outcomes for placebo and nocebo effects were clinical pain reduction after needle intervention and needle-related pain intensity after needle intervention, respectively. While the secondary outcomes were clinical pain intensity before needle intervention, clinical pain intensity during a provocative test after needle intervention, needle-related pain during needle intervention, and days until needle-related pain sensation completely disappears after needle intervention. Outcome measures in this study are detailed in Table 1. Description and clinical findings in VISA-p and ultrasonographic measures after the 8-week period are available elsewhere (27).

Statistical analysis was performed using SPSS, v.25 (IBM Corp., Chicago, IL). A P-value < 0.05 was accepted as a significant difference between compared variables. Variables distribution was assessed using the Shapiro-Wilk test and described in percentage, mean and standard deviation or median and interquartile range, according to the distribution of data. Differences between groups were compared using chi-squared tests (χ²) for categorical data and mixed-model repeated-measures analysis of variance (RM-ANOVA) for continuous data with time (session 1, 2, 3, and 4) as within and group (no-sham group, single-sham group, and double-sham group) as between factors. Pairwise Bonferroni comparisons were performed as post hoc analyses.

The original sample size was calculated based on the VISA-p, and details were reported elsewhere (27). However, a secondary post hoc sample size calculation with G*Power (v3.1.9.2, Heinrich-Heine-University, Dusseldorf, Germany) revealed the feasible sample size for a mixed model RM-ANOVA with three groups (no-sham group, single-sham group, and double-sham group) participating in four experimental sessions. With a power of 90% and an alpha level of 0.01, a total of 42 participants (14 per group) were needed for participation to detect the minimal important difference of 1.2 points (partial η² = 0.05) in the Visual Analogue Scale (28).

No time and group interaction (RM-ANOVA: F_6,132 = 1.1; P = 0.357) was found for pain intensity before needle intervention, indicating no differences in the evolution of pain intensity between groups across sessions. However, a time effect was found (RM-ANOVA: F_3,132 = 16.5; P < 0.001), indicating that the three groups decreased pain intensity over the sessions. Post-hoc analysis showed a lower pain intensity at the beginning of session 4 compared to session one in all groups (no-sham group: P < 0.001; single-sham group: P = 0.011, and double-sham group: P = 0.037). See Table 3.

Significant differences existed between the frequencies of participants reporting pain in the three groups during all the needle interventions (χ² ≥ 10.6; P ≤ 0.005). Specifically, 80% of the no-sham group, 81% to 100% of the single-sham group, and 25% to 44% of the double-sham group reported pain during needle intervention. See Table 3.

No time and group interaction (RM-ANOVA: F_6,132 = 0.7; P = 0.682) or time effect (RM-ANOVA: F_3,132 = 1.7; P = 0.168) was found for pain intensity after needle intervention, indicating that the painful sensation after needle intervention was similar between groups in all sessions. See Table 3.

No time and group interaction (RM-ANOVA: F_6,132 = 0.4; P = 0.937) or time effect (RM-ANOVA: F_3,132 = 1.7; P = 0.179) was found for painful sensation after needle intervention, indicating that the painful sensation after needle intervention was similar between groups in all sessions. See Table 3.

No time and group interaction (RM-ANOVA: F_6,132 = 1.0; P = 0.424) was found for pain reduction after needle intervention, indicating no differences in the evolution of pain reduction between groups across sessions. However, a time effect was found (RM-ANOVA: F_3,132 = 6.3; P < 0.001), indicating that pain reduction within the three groups differed between sessions. Post hoc analysis showed a lower pain reduction at session 2 (P = 0.027) and session 4 (P = 0.001), compared to session 1 in all groups. See Figure 3.

There were no differences between the frequencies of participants reporting the duration of pain sensation after any needle intervention (χ² ≤ 8.8; P ≥ 0.184). See Table 3.

No adverse events were reported other than the previously described needle-related pain itself and the post-puncture pain described above, which were part of the outcome measures.

The three groups similarly improving their pain over time, both at rest (before and after intervention) and during a functional provocative test, might reflect different effects. On the one hand, these findings might only indicate a direct positive effect of eccentric exercise on tendon pain, as previously reported (29). Furthermore, it is unlikely that there was no effect at all from the PNE intervention on pain, as previous studies have found a pain reduction effect in tendinopathies (18). Therefore, it is unlikely that eccentric exercise explains the clinical pain improvements observed in the three groups.

On the other hand, the persistence of patellar tendinopathy symptoms for 18 months in the present study, consistent with previous literature, (30) suggests that non-specific effects such as regression to the mean or natural history might partially explain the observed improvements in clinical pain. However, both the present study and previous literature show a high standard deviation in the duration of symptoms, either at the mean value or above. This substantial variability makes regression to the mean and natural history less likely to account for the clinical pain improvement observed in the three groups. Moreover, it is important to note that these non-specific factors would not explain the observed nocebo effect in needle-related pain. Therefore, it can be inferred that non-specific effects, such as placebo and nocebo responses, are at least partially responsible for the reduction in pain among patients with patellar tendinopathy undergoing needling therapies.

A recent study has shown that expectations may not play as significant role as prior therapeutic experiences in placebo and nocebo responses in healthy individuals (31). Although the connection between these non-specific effects and the psychological characteristics of the participants remains unknown, a recent study with high methodological quality and one of the largest samples in placebo research has shown that individuals with chronic pain benefiting from placebo are characterized by low emotional stress, pain-related fear, and catastrophizing (5). However, catastrophizing has been linked to patellar tendinopathy only in individuals with more severe symptoms (32). The present study, based on the baseline pain VAS values of participants, indicates that in most cases, tendinopathies were moderate to low. Therefore, it can be hypothesized that the sample in the current study did not have a high psychological burden and were more likely to benefit from placebo effects.

As expected, most participants reported pain during the intervention when a needle was inserted into the tendon. Interestingly, up to 44% of participants in the double-sham group also reported pain during intervention. Furthermore, all groups reported the same amount of needle-related pain immediately after the intervention, and up to 25% of participants in the double-sham group continued experiencing pain even 24-48h after the intervention. Overall, these findings indicate that needling interventions are susceptible to non-specific effects, aligning with previous literature (33).

Although not measured on participants, subjective factors such as expectations may have modulated pain (34) after the needling intervention (35). Increased pain due to the expectation of more pain following an intervention is a nocebo effect observed previously (36, 37). A study in patients with chronic shoulder pain found increased mechanical hyperalgesia immediately following sham dry needling that lasted for 24h in individuals with shoulder pain. The study attributed these changes to a nocebo effect generated by negative expectations associated with the instructions given before the procedure (38). In addition to negative expectations, other factors that might have mediated a nocebo effect in post needling-related pain include classical conditioning, observational learning, (39) and prior therapeutic experiences (31).

These findings clearly reflect that needling interventions in individuals with patellar tendinopathy involve intrinsic positive and negative non-specific factors. Therefore, the contextual factors of needling therapies can trigger both placebo and nocebo responses and must be considered in the clinical context. This serves as a general recommendation for physiotherapists and musculoskeletal clinicians, who are encouraged to understand and manage the contextual factors [e.g., patients’ expectations, past treatments, verbal suggestions (40)] that enhance placebo effects and avoid nocebo effects (2, 41).

In this context, it has been shown that individuals with low back pain who received an intense briefing on the adverse side effects of acupuncture tended to exhibit a higher adverse side effect score compared to participants who received a regular adverse side effect briefing (42). However, other studies did not find a significant effect of patient expectations on the short-term effects of DN on pain intensity and pain-inducing stimulus intensities in people with neck pain; (43) or showed that DN treatment can produce beneficial effects on neck pain and tissue mechanosensitivity, regardless of whether participants received a positive, negative, or neutral verbal stimulus before treatment (44).

All participants in the current study experienced post-DN soreness that did not influence treatment outcomes. Interestingly, participants were not very concerned about post-needling soreness as long as their clinical pain complaint decreased. Therefore, it seems that briefings about treatment in routine care might not be as important as previously thought (42). Other variables, such as previous therapeutic experiences (31) or patient satisfaction, might also be important for clinically managing placebo and nocebo effects, as these are factors that can impact the outcome of rehabilitation (45). Indeed, according to several double-blind trials testing pain treatments, the placebo effect can be similar to specific treatment effects (9). In summary, current research on placebo effects indicates that the ethical enhancement of a placebo without using placebos or misinforming patients is feasible and ethical.

This study did not control for demand characteristics, (46) potentially influencing results as participants may have guessed evaluator’s expectations. Including measures of psychological factors such as fear, emotional stress, catastrophizing, and participants’ satisfaction or subjective perception of improvement as indirect measures of the expectations, would have strengthened the study design. This would have allowed ascribing the consistent placebo and nocebo effects found to specific contextual factors. Future studies should explore placebo and nocebo effects in other needle-based techniques for the treatment of musculoskeletal pain.