**3.2 Characteristics of the reviewed articles**

102 articles met the inclusion criteria of the research in the period between 2000 (January) and 2020 (January) in form of original articles, case reports and reviews.

From these 102 articles selected, 23 refer to pneumothorax including more than 120 cases (19 of acupuncture, 3 of injection and 2 of dry needling); 4 articles refer to cardiac tamponade with more than 25 cases (both of acupuncture), 21 in relation to infections, abscesses or hemorrhages (14 of acupuncture, 6 of injection and 1 of dry needling) and other 7 articles refer to adverse effects such as syncope and cardiogenic shock (acupuncture), 3 pneumoperitoneo (acupuncture), 9 hematoma (6 of acupuncture, 1 of injection, 2 of dry needling), hemiplegia (acupuncture), cardiac perforation (acupuncture), hypokalemic paralysis (injection), 6 neuropathies (injection), 1 neuroapraxia (dry needling) and 12 cutaneous lesions/Nicolau syndrome/necrosis (3 of acupuncture, 9 of injection). In 11 articles there already was an existing disease, in other 4 the needle was broken and unfortunately in 4 articles the consequences were fatal. It has also collected 21 review articles of these needling therapies (more than 21.000 adverse effects described). Finally, 6 articles have synthetized information about benefits, risks, perception of security and even modifications of the application of these techniques (**Figure 4**).

**Figure 4.**

*Prevalence of most common adverse events.*

### **4. Discussion**

Considering the outpatient care treatment, adverse effects are possible complications that can occur during or even after the application of these techniques. In more cases there has little importance such as pain, a slight bleeding or a small bruise that disappears quickly. However, other adverse effects without a clear cause can suppose a serious risk for the patient.

These risks have always been present, but in recent years publications have increased considerably. There is no consensus about the classification of these adverse effects. Some authors [111] categorized them into four groups: delayed or missed diagnosis, adverse effects during treatment, bacterial or viral infections, or tissue or organ trauma.

#### **4.1 Incidence/frequency of adverse effects**

The incidence/frequency of these adverse effects is not clear. Acupuncture seems to have an incidence of 2/125.000 cases [30]; White et al. [58] estimated the risk of a serious adverse event with acupuncture at 0.05 per 10.000 treatments, and 0.55 per 10.000 individual patients, Lin et al. [70] reported a pneumothorax incidence of 0.87 per 1.000.000 acupuncture treatments and 1.75 per 1.000.000 in anatomical risk areas; these authors also showed a cellulitis incidence [72] about 64.4 per 100.000 treatments.

In relation to injections, Anderson et al. [97] explained an incidence of 5.8% of adverse effects. Kim et al. [93] introduced the incidence separately in relation with the case: 0.84% and 1.63% in relation to the patient; on the other hand, the procedure had an incidence of 0.07% and the administrated drug 0.15%. Finally, the unknown etiology had a 0.63% for this author [93] and for other authors it is unknown [121].

Data about incidence of dry needling procedures has not been found.

The most reviewed articles refer isolated cases and not a periodicity, but other authors have published several reviews that try to synthesize this information. Considering these 3 needling techniques, acupuncture leaves a clear superiority in relation to the number of publications with adverse effects.

Peuker et al. [20] investigated the traumatic wounds caused by acupuncture and discuss how these complications could be avoided. Lao et al. [22] reviewed 98 publications (1965–1999) and they found 202 complications (infections, tissue/ organ damage and nerve injury). Cutaneous disorders, hypotension, fainting and vomiting were some adverse effects described. Chauffe et al. [29] found 9 cases of pneumothorax since 1985. Witt et al. [36] reviewed acupuncture studies in chronic osteoarthritis pain of the knee or hip, lumbar, cervical, head, allergic rhinitis, dysmenorrhea and asthma. Out of 229.230, 19.726 reported at least 1 adverse effect (bleeding, pain, vegetative symptoms). The longest duration of these adverse effects was 180 days (nerve injury). Ernst et al. [39] conducted a review of cardiac tamponade after acupuncture: 26 cases were found and 14 with fatal complications. He et al. [43] reviewed 167 articles with 1.038 cases (35 deaths) from Chinese literature. 468 cases were syncope, 307 pneumothorax, and 64 subarachnoid hemorrhage. Xu et al. [44] checked the frequency and severity of these effects (2000–2011): 117 articles with 308 adverse effects in 25 countries (294 for acupuncture, 4 moxibustion and 10 cupping). Peuker et al. [52] reviewed the traumatic lesions after acupuncture. Wu et al. [53] performed a review in China (1980–2013), finding 182 incidents in 133 papers (internal organ, tissue and nerve injury are the major complications). The adverse effects included were syncope, infection, hemorrhage, allergy, burn, aphonia, hysteria, cough, thirst, fever, somnolence and broken needles. White et al. [58] found 715 incidents in their review: 90 reports of trauma and 12 reports of death. In Taiwan, Lin et al. [70, 72] published 2 reviews (1997–2012) about pneumothorax and cellulitis incidence respectively. They evaluated 411.734 patients with 5.407.378 treatments of acupuncture [70] and 407.802 patients with 6.207.378 acupuncture treatments [72]. In both articles the authors emphasized the importance of the previous medical history. Ullah et al. [77] reviewed 133 articles and selected 30 cases with relevant cardiac complications: 8 were infective complications and 22 cardiac tamponades.

Regarding injections, 8 articles have been found. Antoni Ro et al. [79] reviewed 226 cases (1955–1959), finding 71 cases of pneumothorax and Cheng et al. [81] performed a review (1966–2006) explaining the complications of this technique: "infections, spinal cord injury and peripheral nerve injuries, pneumothorax, air embolism, pain or swelling at the site of injection, chemical meningism, granulomatous inflammation of the synovium, aseptic acute arthritis, embolia cutis medicamentosa, skeletal muscle toxicity, and tendon and fascial ruptures". De la Torre et al. [87] introduced a review about the adverse effects caused by botulinum toxin A in masticatory muscles. They used 436 citations and concluded with 16:7 were myofascial pain and 9 were trigeminal neuralgia. The most frequent adverse effects were "temporary regional weakness, tenderness over the injection sites and minor discomfort during chewing". Most of them had a spontaneous resolution. Marcus et al. [91] found a very rare complication due to injections (diclofenac, dexamethasone and benzathine penicilin): Nicolau Syndrome. They found 4 cases (2016–2018). Park et al. [93] investigates an association between steroid injection and calcification in lateral epicondylitis. They evaluated 110 patients (February 2016-October 2018) and concluded that the injections history and the number of them has a significative association with soft tissue calcifications. A review (January 2000-April 2016) about adverse events due to intra-articular corticosteroid injections was made by Anderson et al. [97] 1.708 patients from 3 regional hospital participated: 99 patients had 104 adverse effects within 90 days post-injection. The most prevalent symptom was flare (78 patients) and 10 patients had skin reactions. There were no infections. Years before, Kim et al. [98] had reviewed 11.980 injections in 6.066 patients (January 2007-December 2017). There were 101 facet-joint injections and 99 patients developed adverse effects. 7 patients had an infectious spondylitis, 1 patient died of an uncontrolled infection and 2 patients had partial recovery of their neurological

condition. Finally, Desai et al. [107] published a review of 17 years where reflected the iatrogenic peripheral nerve injuries due to injections. They included "intramuscular injections, brachial nerves procedures, subclavian and jugular venous cannulation and routine intravenous injections". The most frequents symptoms were pain, paresthesia and sensory-motor deficits. 190 patients needed surgical intervention, 164 had any sequel or no recovery and 9 had neurological deterioration with weakness.

There is not standard data on the incidence of these events. Unfortunately, the huge diversity of pathologies, interventions, therapists… makes difficult a generalization.

#### **4.2 Most prevalent adverse effects**

There is no consensus about the most frequent adverse effect in the literature. Some of them are pneumothorax, cardiac tamponade, air embolism, spinal epidural haematoma/abscess, abdominal visceral injury, median and fibular nerve injury and infection [20, 36, 75, 77, 81, 87, 111].

Some authors reflected that pneumothorax is the most cited adverse effect, [50, 63] while for others is infection [44]. White et al. [58] agree with both theories being the most common complication pneumothorax and injury to the central nervous system and infection will be in second place. Ullah et al. [77] concluded that cardiac tamponade is the most frequent complication.

It seems that invasive techniques on the thorax are related to a high incidence of pneumothorax [118]. There are some investigations in different countries (United Kingdom, Japan, Czechoslovakia, Switzerland, Germany, Japan and Taiwan) about it. The incidence of these cases is low, less than 1/10.000. However, there have been more than 100 cases reporting iatrogenic pneumothorax due to acupuncture and dry needling, including cases of death [112]. Lin et al. [70] showed an incidence of 0.84/1.000.000 and 1,75/1.000.000 at risk anatomical areas. Most iatrogenic pneumothorax used to be unilateral, but there are bilateral cases too [26, 30]. In this article there are 23 articles related to pneumothorax [19, 26, 29, 30, 33, 46, 47, 49, 50, 55, 56, 59, 62, 63, 68–70, 78–80, 83, 112, 118].

Other incidents (less frequent) reported in the literature but not less important are cardiac tamponade [21, 39, 51, 75, 77], granulomas [23], fistulas [24, 38], necrosis [25, 42, 85], infections [27, 57, 119], abscesses [27, 45, 61, 71, 74, 76], pneumoretroperitoneum [32], hemorrhages [35], hemopericardium [37], haematomas [40, 64, 67, 109, 111, 117], chilotorax [41], organ perforation [48, 60], needle rupture [53, 55, 117], hemiplegia [54], hemothorax [56], peritonitis [66], cellulitis [72], hypokalemic paralysis [84], nerve injury [86, 99, 102, 107, 108, 110, 116], weakness [88], necrosis [89, 90, 94, 103], Nicolau Sydrome [91, 96, 101], toxicity…[92] Almost all had a complete resolution of the symptoms. However, publications with fatal and irreversible consequences have also been found [24, 37, 98].

#### **4.3 Special considerations**

There are several aspects must be considered when carrying out these techniques in the treatment of muscular pain in outpatients. These incidents, even taking caution may occur; therefore, it is important to obtain a complete clinical history highlighting possible underlying pathologies [70–72]. Several articles have found patients with asthma [19], diabetes [25, 85], anemia [42], herpes zoster [47], cancer [51], miastenia gravis [55] and scoliosis [57] and sclerosis [89]. These pathologies could influence the appearance or greater probability of developing a complication.
