|
 
 |
| ORIGINAL ARTICLE |
|
| Year : 2020 | Volume
: 37
| Issue : 2 | Page : 94-99 |
|
Bilateral nerve conduction studies must be considered in the diagnosis of sciatic nerve injury due to intramuscular injection
Halit Fidanci1, İlker Öztürk2, Ahmet Candan Köylüoğlu2, Mehmet Yıldız2, Zülfikar Arlıer2
1 Department of Neurology, Division of Clinical Neurophysiology, Adana City Training and Research Hospital; Department of Neurology, Adana City Training and Research Hospital, Adana, Turkey
2 Department of Neurology, Adana City Training and Research Hospital, Adana, Turkey
| Date of Submission | 05-Sep-2019 |
| Date of Decision | 18-Nov-2019 |
| Date of Acceptance | 22-Nov-2019 |
| Date of Web Publication | 29-Jun-2020 |
Correspondence Address:
Halit Fidanci
Department of Neurology, Division of Clinical Neurophysiology, Adana City Training and Research Hospital, Yuregir 01060, Adana
Turkey
 Source of Support: None, Conflict of Interest: None  | 4 |
DOI: 10.4103/NSN.NSN_22_20
Objectives: Although compound muscle action potential (CMAP) and sensory nerve action potential (SNAP) amplitudes of the nerves are reduced in sciatic nerve injury due to intramuscular injection (SNIII), they may still be higher than the reference values if there is a mild axonal degeneration. In this case, comparing the outcomes of nerve conduction studies of intact and affected lower extremities becomes important. We aimed to determine the role of this comparison in the diagnosis of SNIII. Methods: Patients with SNIII were included. Reference values for lower extremity nerve conduction studies were obtained from healthy participants. Peroneal, posterior tibial, superficial peroneal, and sural nerve conduction studies were performed in both lower extremities. In the first analysis, the CMAP or SNAP amplitude of the nerve was considered abnormal if it was less than the reference value. In the second analysis, the CMAP or SNAP amplitude of the nerve was considered abnormal if it was less than the reference value or <50% of the CMAP or SNAP amplitude obtained from the intact limb nerve. Results: Thirty patients and 31 controls were included in the study. Compared with those found in the first analysis, the number of posterior tibial nerve CMAPs with reduced amplitudes, and the sural and superficial peroneal nerve SNAPs with reduced amplitudes were higher in the second analysis (P = 0.008, P < 0.001, and P = 0.031; respectively). Conclusion: This study showed that nerve conduction studies should be performed in both the intact and affected extremities in SNIII.
Keywords: Electrodiagnosis, nerve conduction study, sciatic nerve injury
How to cite this article:
Fidanci H, Öztürk &, Köylüoğlu AC, Yıldız M, Arlıer Z. Bilateral nerve conduction studies must be considered in the diagnosis of sciatic nerve injury due to intramuscular injection. Neurol Sci Neurophysiol 2020;37:94-9 |
How to cite this URL:
Fidanci H, Öztürk &, Köylüoğlu AC, Yıldız M, Arlıer Z. Bilateral nerve conduction studies must be considered in the diagnosis of sciatic nerve injury due to intramuscular injection. Neurol Sci Neurophysiol [serial online] 2020 [cited 2023 Jun 10];37:94-9. Available from: http://www.nsnjournal.org/text.asp?2020/37/2/94/288418 |
| Introduction | |  |
Sciatic nerve injury (SNI) may develop as a result of intramuscular (IM) injection.[1],[2],[3] Although hip arthroplasty is known to be the most common cause of SNI, several studies have reported IM injections as the leading cause in poor economies.[1],[2],[3],[4] Age, weight, needle length, the position of the patient at the time of injection, type of intramuscularly administered agent, and the anatomic variations of the muscles or nerves are the major risk factors for sciatic nerve injury due to IM injections (SNIIIs).[1],[3],[5],[6],[7] The diagnosis of SNIII can be made clinically and by using the findings from neurologic examinations, nerve conduction studies, electromyography (EMG), and magnetic resonance imaging.[1] Patients with SNI have weakness in the muscles innervated by the peroneal and/or tibial and/or sciatic nerves, with paresthesia in the respective dermatomes. Nerve conduction studies and needle EMG have an important role in the diagnosis and differential diagnosis of SNI.
Axonal degeneration is predominant in SNIII.[1],[4],[8] Due to axonal damage, the amplitudes of compound muscle action potentials (CMAPs) and sensory nerve action potentials (SNAP) are reduced. However, in cases of mild axonal damage, and especially in younger individuals, CMAP and SNAP amplitudes can be found to be higher than reference values even if they are reduced.[9],[10] In this case, the side-to-side comparison of nerve conduction studies becomes important. Therefore, we aimed to investigate the role of comparing nerve conduction study findings obtained from the affected and unaffected extremities in the diagnosis of SNIII.
| Methods | | |
Subjects
Patients with clinical features and electrodiagnostic test findings compatible with SNIII were included in this retrospective cohort study. The control group consisted of healthy individuals. Individuals with polyneuropathy, mononeuropathy, neurodegenerative diseases, and diseases that may cause polyneuropathy such as diabetes mellitus, and individuals with paresthesia or weakness in the extremities were not included in the control group. The reference values for nerve conduction studies were obtained from the controls. The clinical and electrodiagnostic findings of the patients who presented to EMG Laboratory of Adana City Research and Training Hospital (ACRTH) between July 2018 and July 2019 were retrospectively analyzed. Patients were included in the study if they met the following criteria:(1) following the IM injection, symptoms should have begun immediately or within minutes or hours; (2) weakness present in at least one of the muscles innervated by the peroneal or tibial or sciatic nerves; (3) needle EMG findings consistent with SNI. The exclusion criteria were polyneuropathy or a disease that may cause polyneuropathy such as diabetes mellitus; clinical and electrodiagnostic test findings suggestive of lumbosacral radiculopathy or plexopathy or a mononeuropathy such as peroneal neuropathy; a history of lumbosacral radiculopathy or plexopathy; and neurodegenerative diseases. The Leeds Assessment of Neuropathic Symptoms and Signs (LANSS) was used to evaluate the severity of neuropathic symptoms.[11] Ethics committee approval was obtained from the ethics committee of ACRTH (number: 37/515). Written informed consent was obtained from all participants.
Electrodiagnostic tests
The nerve conduction studies and needle EMG were performed using a Cadwell Sierra Summit EMG unit (Cadwell Laboratories, Kennewick, Washington, USA). The processes of stimulation and recording were performed using surface electrodes. Motor unit action potentials (MUP) and the presence of active denervation were analyzed using concentric needle EMG electrodes (length = 50 mm, diameter = 0.46 mm, Bionen Medical Devices, Florence, Italy). The electrodiagnostic tests were performed when the skin temperature was above 32°C. Extremities with lower temperatures were warmed appropriately to perform the electrodiagnostic tests. For the sensory and motor nerve conduction studies and the needle EMG, the low-high frequency filters were set to 20 Hz–2 kHz, 20 Hz–10 kHz, and 10 Hz–10 kHz, respectively. The sweep speed was 5 ms and 1 ms/division for the motor and sensory nerve conduction studies, respectively. It was 10 ms, 10 ms, and 100 ms/division to analyze active denervation, MUP, and the recruitment pattern, respectively. The sensitivity for sensory and motor nerve conduction studies, the spontaneous activity, and MUP analyses were set to 10 μV, 5 mV, 50–100 μV, and 200–1000 μV/division, respectively. Peak-to-peak CMAP and SNAP amplitudes were determined. The sensory nerve conduction studies were performed antidromically. The sural nerve conduction velocity was calculated using peak latency, and superficial peroneal nerve conduction velocity was calculated using onset latency. The peroneal nerve CMAP was recorded from the extensor digitorum brevis and tibialis anterior muscles. The posterior tibial nerve CMAP was obtained from the abductor hallucis muscle. To obtain distal CMAPs, the peroneal nerve from the extensor digitorum brevis muscle and the posterior tibial nerve were stimulated at the ankle 8 and 10 cm proximal to the recording electrode, respectively. In the first analysis, the amplitude of CMAP or SNAP was considered abnormal when it was less than the reference value. In the second analysis, when the amplitude of CMAP or SNAP was less than the reference value or <50% of the amplitude of the intact limb nerve CMAP or SNAP, it was considered abnormal. Needle EMG was performed visually. The presence of active denervation was carefully evaluated. According to the tolerability level of the patient, 10–20 MUPs were analyzed during mild contractions in each muscle. MUP was considered neurogenic if the MUP amplitude was >4 mV and/or its duration was >15 ms. The SNI protocol implemented in our EMG laboratory involved applying needle EMG to the tibialis anterior, medial gastrocnemius, peroneus longus, short head of biceps femoris, vastus lateralis, gluteus maximus, gluteus medius, and the L3, L4, L5, S1 paraspinal muscles. The number of muscles examined could vary according to the tolerability level of the patient. When the diagnosis of SNIII was suspected, needle EMG was performed bilaterally in the lower extremities.
Statistical analyses
The distribution of the variables was checked using the Shapiro–Wilk test. The Mann–Whitney U-test, Student's t-test, and Wilcoxon's signed-ranked test were used in the analysis of quantitative data, and the Chi-square test and McNemar test were used in the analysis of qualitative data. Mean ± standard deviation (SD) and the median of numeric data were calculated for descriptive statistics. The upper and lower limits were calculated as mean ± 2 SD for normally distributed variables and as 2.25th or 97.75th percentile values for data that were not normally distributed. P < 0.05 was considered statistically significant. The Statistical Package for the Social Sciences (SPSS IBM Corp., Armonk, NY, USA) version 22.0 was used to perform the statistical analysis.
| Results | | |
Thirty patients with SNIII and 31 controls were included in the study. The demographic characteristics of the controls and patients are shown in [Table 1]. One patient had diabetes mellitus and one patient had a history of lumbosacral radiculopathy; these two patients were excluded from the study. Age, sex, and height were not different between the two groups. Weight and body mass index (BMI) were significantly lower in the patient group compared with the control group [P < 0.001, [Table 1]. The BMI was <18.5 kg/m 2 in 11 patients. The neurologic examination findings are shown in [Table 1]. IM injections were found to be administered to the dorsogluteal area by paramedics or nurses in all patients. The left sciatic nerve was affected in 20 patients. Sensory loss was most commonly present in the dorsum of the foot. All patients had weakness in at least one of the muscles innervated by the peroneal nerve. Two patients had a decreased Achilles reflex, and 28 patients had no Achilles reflex. LANSS was performed in 28 patients. The mean LANSS score was 13.1 ± 5.9 (range, 3–24). The LANSS score was ≥12 in 17 patients. The mean time from the IM gluteal injection to the electrodiagnostic test was 7.7 ± 7.0 (range, 1–24) months. This interval was <1 year in 24 patients (80%). Symptoms were found to have occurred immediately after the IM injection in 26 patients and within minutes or hours in four patients. The number of patients who developed SNIII due to analgesics, antibiotics, analgesics + antibiotics, analgesics + muscle relaxants, and anti-allergy drugs was 20, 3, 2, 1, and 1, respectively. Diclofenac caused SNIII in 17 patients. The agent causing SNIII was unknown in three patients. Eleven patients had received IM injections due to fever or infections. Other indications for the IM injections were found as joint pain, headache, toothache, nephrolithiasis, abdominal pain, generalized pain, and allergy. There were no patients with low back pain. [Table 2] shows the reference values of posterior tibial, peroneal, superficial peroneal, and sural nerve conduction studies obtained from the controls. | Table 2: Reference values of posterior tibial, peroneal, superficial peroneal, and sural nerve conduction studies obtained from controls
Click here to view |
One hundred fifty CMAPs and SNAPs of four nerves from 30 patients were examined. The amplitude abnormalities of CMAPs and SNAPs from the patients obtained in the first and second analyses are shown in [Table 3]. In the first and second analyses, the results of the sensory nerve conduction studies were normal in eight and three patients, respectively. The sural and superficial peroneal nerve SNAP amplitudes were reduced in nine and 16 patients according to the reference values, respectively. In the second analysis described in the method section of the article, abnormalities of the sural and superficial peroneal nerve SNAP amplitudes were detected in 23 and 22 of 30 patients, respectively. The number of patients with reduced sural and superficial nerve SNAP amplitudes was significantly higher in the second analysis than in the first analysis (P < 0.001, P = 0.031). A similar finding was also present in the CMAP amplitude of the posterior tibial nerve (P = 0.008). In the first analysis, amplitudes of CMAPs and SNAPs were normal in all nerves in six patients, whereas in the second analysis, five of these six patients had abnormalities in at least two CMAPs or SNAPs. One patient had normal nerve conduction studies in the first and second analyses, although there was mild weakness of foot dorsiflexion and abnormal needle EMG findings in the tibialis anterior and peroneus longus muscles. The number of patients with reduced amplitude in at least two and three CMAPs or SNAPs was significantly higher in the second analysis than in the first analysis [P = 0.016, P = 0.002, [Table 3]. In the first analysis, the amplitudes of CMAPs or SNAPs were found to be reduced in 61 (41%) of the 150 CMAPs and SNAPs examined in nerve conduction studies, whereas in the second analysis, the number of CMAPs or SNAPs with abnormal amplitude was 98 (65%). The number of abnormal CMAP and SNAPs in the second analysis was significantly higher than in the first analysis (P < 0.001). Furthermore, [Figure 1] shows that the number of abnormal CMAPs or SNAPs in each patient was significantly higher in the second analysis (P < 0.001). Active denervation findings or neurogenic changes were present in the tibialis anterior and medial gastrocnemius muscles of 22 (73%) and 19 (63%) of the 30 patients, respectively. Needle EMG was performed to the peroneus longus and short head of biceps femoris muscles of 28 patients. There were active denervation findings or neurogenic changes in 16 (57%) peroneus longus and 19 (68%) short head of biceps femoris muscles. No clinical and electrodiagnostic findings suggestive of lumbosacral radiculopathy, lumbosacral plexopathy or mononeuropathy such as peroneal neuropathy were found in the patients. | Table 3: Comparison of nerve abnormalities based on the first and second analyses
Click here to view |
 | Figure 1: Comparison of the number of abnormal compound muscle action potentials and sensory nerve action potentials amplitudes obtained from patients in the first and second analyses. In the first and second analyses, the mean of number of compound muscle action potentials and sensory nerve action potentials with abnormal amplitudes were 2.1 ± 1.3 (median = 2, range 0–4) and 3.3 ± 1.2 (median = 3, range 0–5), respectively. The mean number of compound muscle action potentials and sensory nerve action potentials with abnormal amplitude was statistically higher in the second analysis than in the first analysis (P < 0.001). In one patient, the first and second analyses revealed no abnormalities in the nerve conduction studies
Click here to view |
| Discussion | | |
SNIII is an important health problem because it can cause disability in patients. Direct needle trauma, reduced protective tissue in the gluteal area, angle or length of the needle, and neuronal ischemia caused by scar formation due to drug injection may cause SNIII.[1],[3],[5],[12] Thin individuals in particular are known to be more susceptible to SNI because of the small amount of gluteal protective tissue.[1],[2],[5] A supporting finding was also present in our study that the BMI and body weights of the patients were significantly lower than those of controls. The BMI was <18.5 kg/m 2 in 11 patients. For these reasons, factors that may cause SNIII such as the position of the patient or the needle length, should be taken into consideration when an IM injection is given to individuals with low BMI. In these patients, the ventrogluteal area may be preferred for IM injection instead of dorsogluteal area. The dorsogluteal area was used for IM injections in all patients in our study. There are reports that the ventrogluteal area is safer for IM injections than the dorsogluteal area.[1],[7] Consistent with the literature, weakness was more common in the muscles innervated by the peroneal nerve in our study.[1],[2],[4] The absence or reduction of the Achilles reflex in all patients indicates the importance of the this reflex in SNI. Although the patients reported weakness or pain rather than sensory symptoms, our neurologic examinations revealed that most patients also had sensory loss on the dorsum of the foot. The LANSS score was ≥12 in 17 patients, indicating that pain was an important symptoms.[1],[5] Similar to the literature, in most of the patients (87%), symptoms began immediately after the IM injection, which is thought to be due to intraneural injection.[1],[5] There were also patients (13%) whose symptoms began minutes or hours after IM injection, possibly due to diffusion after injection into the epineurium or close to the nerve.[1]
When the results of the first and second analyses were compared, the number of CMAPs or SNAPs with reduced amplitudes was found to be significantly higher in the second analysis than in the first analysis. This difference can be explained by the presence of mild axonal degeneration in some patients. Histologic and electrodiagnostic studies have shown that axonal degeneration is prominent in SNIII.[1],[4],[8] The severity of this degeneration may vary from mild to severe.[1],[13] Because basal CMAP and SNAP amplitudes are high in some individuals, especially young individuals, they may still be found within the normal limits despite reductions due to mild axonal degeneration.[9],[10]
The number of sural, posterior tibial, and superficial peroneal nerves with abnormal CMAP or SNAP amplitudes were significantly higher in the second analysis when the results from the first and second analyses were compared. This difference was more prominent in the posterior tibial and sural nerves. The number of peroneal nerves with abnormal CMAP amplitudes was not different between the two analyses. These findings were not surprising because the peroneal division of the sciatic nerve is affected more extensively compared with the tibial division of the sciatic nerve in SNIII.[1],[7],[14] This may occur because the peroneal branch is located more laterally and there is less protective tissue in this area.[14],[15] For these reasons, axonal degeneration will be more pronounced in the peroneal division of the sciatic nerve and the CMAP amplitude of the peroneal nerve may be lower than the reference values. The presence of CMAP and SNAP amplitude abnormalities, found mostly in the peroneal and superficial peroneal nerves in the first analysis, supports this conclusion.
Performing nerve conduction studies in the intact extremity increases the sensitivity when diagnosing SNIII. In our study, nerve conduction studies of six patients were normal. However, when the results were re-evaluated in respect to both the reference values and the nerve conduction study findings obtained from the intact extremities, the CMAP or SNAP amplitudes were found to be abnormal in at least two CMAPs or SNAPs of five out of these six patients. Furthermore, the number of patients with two and three CMAPs or SNAPs with reduced amplitudes was lower in the first analysis than in the second analysis. All these findings suggest that nerve conduction studies should be performed in both lower extremities in SNIII.
Conventionally, nerve conduction studies are performed in each lower extremity in some EMG laboratories. Our study has shown that this is necessary. Due to the high number of patients admitted to some EMG laboratories, it may not be possible to perform bilateral nerve conduction studies of all nerves. In this case, our study showed that it might be sufficient to perform sural and posterior tibial nerve conduction studies bilaterally to save time. To the best of our knowledge, there is only one electrodiagnostic study in the literature comparing side-to-side nerve conduction studies in SNI.[4] However, that study examined patients with SNIII and patients with SNI due to other causes. In that study, sciatic nerve SNAP was normal in 29% of patients, side to side nerve conduction studies were not applied in all patients. In our study, the sural nerve SNAP was normal in 23% of patients, and this rate was slightly lower than the rate reported by Yuen et al. This lower rate might have occurred because we performed the nerve conduction studies in both lower extremities in all patients. Both sural and superficial nerve SNAPs were normal in 10% of the patients, consistent with the literature.[4] A sural or superficial peroneal nerve SNAP abnormality occurs in the majority of patients with SNIII, but finding normal SNAP amplitudes in these nerves alone does not exclude the diagnosis of SNI. In our study, a normal nerve conduction study in one patient supports this assertion. The presence of needle EMG abnormalities in at least one muscle in all patients indicates the importance of needle EMG in the diagnosis of SNIII. Similar to previous studies, needle EMG abnormalities were found to be higher in peroneal innervated muscles.[4]
There were some limitations in our study. First, although the interval between the time of IM injection and electrodiagnostic tests was <1 year in the majority of patients, this interval ranged from 1 month to 24 months. Second, we included patients with muscle weaknesses in our study, which may be another limitation. It may be interesting to compare the findings from nerve conduction studies performed in intact limbs and involved limbs that cause symptoms in patients without muscle weaknesses. Finally, the retrospective nature of the study was the other limitation of the study.
| Conclusion | | |
This study showed that nerve conduction studies should be performed in the affected extremity and the intact extremity in SNIII. When the CMAP or SNAP amplitude of the nerve is less than the reference value or <50% of the CMAP or SNAP amplitude of the intact extremity nerve, we recommend that this finding should be accepted as abnormal.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
| References | | |
| 1. | Jung Kim H, Hyun Park S. Sciatic nerve injection injury. J Int Med Res 2014;42:887-97.  |
| 2. | Tak SR, Dar GN, Halwai MA, Mir MR. Postinjection nerve injuries in Kashmir: A menace overlooked. J Res Med Sci 2008;13:244-7. |
| 3. | Senes FM, Campus R, Becchetti F, Catena N. Sciatic nerve injection palsy in the child: Early microsurgical treatment and long-term results. Microsurgery 2009;29:443-8. |
| 4. | Yuen EC, So YT, Olney RK. The electrophysiologic features of sciatic neuropathy in 100 patients. Muscle Nerve 1995;18:414-20. |
| 5. | Yeremeyeva E, Kline DG, Kim DH. Iatrogenic sciatic nerve injuries at buttock and thigh levels: The Louisiana state university experience review. Neurosurgery 2009;65:A63-6. |
| 6. | Streib EW, Sun SF. Injection injury of the sciatic nerve: Unusual anatomic distribution of nerve damage. Eur Neurol 1981;20:481-4. |
| 7. | Mishra P, Stringer MD. Sciatic nerve injury from intramuscular injection: A persistent and global problem. Int J Clin Pract 2010;64:1573-9. |
| 8. | Gentili F, Hudson AR, Kline D, Hunter D. Early changes following injection injury of peripheral nerves. Can J Surg 1980;23:177-82. |
| 9. | Preston DC, Shapiro BE. Electromyography and Neuromuscular Disorders Clinical-Electrophysiological Correlations. 3 rd ed. London: Elsevier Saunders; 2013. p. 29-30. |
| 10. | Preston DC, Shapiro BE. Electromyography and Neuromuscular Disorders Clinical-Electrophysiological Correlations. 3 rd ed. London: Elsevier Saunders; 2013. p. 389-90. |
| 11. | Yucel A, Senocak M, Kocasoy Orhan E, Cimen A, Ertas M. Results of the Leeds assessment of neuropathic symptoms and signs pain scale in Turkey: A validation study. J Pain 2004;5:427-32. |
| 12. | Small SP. Preventing sciatic nerve injury from intramuscular injections: Literature review. J Adv Nurs 2004;47:287-96. |
| 13. | Selander D, Brattsand R, Lundborg G, Nordborg C, Olsson Y. Local anesthetics: Importance of mode of application, concentration and adrenaline for the appearance of nerve lesions. An experimental study of axonal degeneration and barrier damage after intrafascicular injection or topical application of bupivacaine (Marcain). Acta Anaesthesiol Scand 1979;23:127-36. |
| 14. | Pham M, Wessig C, Brinkhoff J, Reiners K, Stoll G, Bendszus M. MR neurography of sciatic nerve injection injury. J Neurol 2011;258:1120-5. |
| 15. | Kline DG, Kim D, Midha R, Harsh C, Tiel R. Management and results of sciatic nerve injuries: A 24-year experience. J Neurosurg 1998;89:13-23. |
[Figure 1]
[Table 1], [Table 2], [Table 3]
| This article has been cited by | | 1 |
ANXIETY AND DEPRESSION IN TYPE 2 DIABETES MELLITUS: A STUDY INCLUDING NERVE CONDUCTION STUDY AND NEUROPATHIC PAIN |
|
| Sencan BUTURAK, Halit FIDANCI, Mehmet YILDIZ, Zülfikar ARLIER | | Kirikkale Üniversitesi Tip Fakültesi Dergisi. 2023; 25(1): 76 | | [Pubmed] | [DOI] | | | 2 |
Neuropathic pain in peroneal nerve entrapment at the fibular head |
|
| Ilker Öztürk, Halit Fidanci, Zülfikar Arlier | | Arquivos de Neuro-Psiquiatria. 2022; 80(11): 1134 | | [Pubmed] | [DOI] | | | 3 |
The Relationship Between Body Mass Index and Mononeuropathies |
|
| Ayse Begüm BÜYÜKSURAL, Halit FIDANCI, Sencan BUTURAK, Ilker ÖZTÜRK, Mehmet YILDIZ, Izzet FIDANCI, Zülfikar ARLIER | | Düzce Tip Fakültesi Dergisi. 2022; | | [Pubmed] | [DOI] | | | 4 |
The relationship between nerve conduction studies and neuropathic pain in sciatic nerve injury due to intramuscular injection |
|
| Halit Fidanci,Ilker Öztürk | | The Korean Journal of Pain. 2021; 34(1): 124 | | [Pubmed] | [DOI] | |
|
 |
|
|
|
Search Pubmed for