The cumulative lifetime prevalence of low-back pain lasting more than 2 weeks is 13.8%.1 Lumbar discectomy has been shown to produce better clinical outcomes than chemonucleolysis or placebo;2 however, a third of patients continue to suffer from persistent pain postoperatively.3–5 Pain and disability at 2 months postoperatively predict those at 12 months postoperatively.6
It has been shown that one-third of patients with radicular pain display characteristics of neuropathic pain7 and reorganization of the primary cortex has been seen in patients with chronic back pain.8 However, it has also been demonstrated that patients with back pain have elevated thresholds to warmth, cold, vibration and electrical stimuli,9–11 and that greater preoperative warmth thresholds predict a poor outcome after surgical decompression.12 Herniated lumbar disc tissue shows a greater production of nitric oxide than control tissue,13,14 and nitric oxide is produced by cells in granulation tissue around herniated disc tissue.15 High levels of synovial (group II) PLA2 have been identified in lumbar disc tissue16 and human disc PLA2 promotes inflammation, sensitization of nerve fibres and neurotoxicity.17–23
Our main hypothesis was that an association exists between patient outcome, defined using the Modified Stauffer–Coventry evaluating criteria,24 and the preoperative pain perception threshold to an electrical stimulus. Our secondary hypotheses were that patient outcome was associated with changes in the pain perception thresholds to an electrical stimulus in the perioperative period, perioperative sensation perception threshold, perioperative serum nitric oxide metabolites, perioperative pain scores and analgesic requirements, mood, and the histological appearance of, and the presence of PLA2 in, excised disc tissue. A pilot study was devised to test these hypotheses.
With institutional ethics approval and having obtained informed consent from each, 43 ASA (American Society of Anaesthesiologists) class I or II patients, with radiological evidence of intervertebral disc herniation at one level and scheduled for lumbar discectomy, were recruited. Exclusion criteria were previous surgery or epidural injection for back pain, allergy to medications contained in the study protocol, neurological disorders (such as multiple sclerosis), diabetes mellitus, other chronic pain syndromes, and patient refusal. Each patient's age, sex, weight, height, the duration of pain preoperatively, involvement in litigation because of back pain, employment status, smoking status, the disc level involved, the side (left or right) and site of pain (leg, back, or leg and back) were noted. Preoperative root tension signs (straight leg raising) were documented. Patients were assessed preoperatively, 4 h, 24 h, and 2 months postoperatively.
Quantitative sensory testing was performed at each of these assessment time points. This consisted of determining thresholds to transcutaneous constant current electrical stimulation using a Keypoint 3.0 system (Dantec Medtronic Int., Denmark). A square wave of 5 μV with a duration of 0.1 ms at a frequency of 1 Hz was used, with a ramping rate of 1 mA/s. With the subject lying supine, the sensation perception and pain perception thresholds were established bilaterally on a predetermined area over the affected dermatome and over the mid-humeral point on the lateral aspect of the contralateral arm (C5). The mean of three measurements, taken 5 min apart, was taken to represent each site. Readings were repeated until a total of three measurements were obtained such that the greater was no more than 120% of the least. The ratio of the pain perception threshold at the L4 or L5 dermatome to that at the control dermatome (i.e. PTT at C5 on the contralateral side) [(L4 or L5 –C5)/C5×100] was calculated.
Four investigators trained in the technique collected the quantitative sensory testing data in this study, and a group of 10 volunteers was recruited to establish interinvestigator variability. Quantitative sensory testing was performed by two investigators over the C5 dermatome as outlined earlier in a double-blind fashion with a rest period of 5 min between the two sets of readings. A random number generator was used to determine the order of investigators and Bland–Altman plots were charted (Medcalc version 7.2, MedCalc Software bvba, Broekstraat 52, 9030 Mariakerke, Belgium).
Venous blood samples (4 ml) were taken for determination of nitric oxide metabolites (nitrate and nitrite) prior to induction, at the end of surgery, and 4 and 18 h postoperatively. Plasma concentrations of stable nitric oxide metabolites were measured using the Nitric Oxide Chemiluminescence Analyser (Sievers 280 NOA, Sievers Instruments, Boulder, Colorado, USA).25
Patients received diclofenac (100 mg) and paracetamol (1 g) rectally, and morphine sulphate 0.1 mg/kg intravenously intraoperatively. Each received regular diclofenac (100 mg 18 hourly) and paracetamol (1 g 6 hourly) ‘around the clock’, and morphine sulphate (0.1 mg/kg 4 hourly) as required postoperatively for 24 h. Analgesic consumption, adverse effects of analgesics (nausea, vomiting, pruritus) and sedation score (0, alert; 1, drowsy; 2, asleep but easily rousable; 3, unrousable) were recorded at 4 and 24 h postoperatively. Visual analogue scales (0–100) for anxiety and pain were completed at each of the time points. The HADS score and the McGill Pain Questionnaire (short form)26 were completed preoperatively and 2 months postoperatively.
At the time of surgery, the surgeon noted whether or not disc tissue was extruded. The excised disc material removed was placed immediately in liquid nitrogen and stored at −70°C. The tissue was formalin fixed and paraffin processed, and a section taken from each paraffin block for staining with haematoxylin and eosin (H and E). A semi-quantitative three-point scoring system [0, no change; 1, degenerative changes (fibrillation, disintegration); 2, necrosis (dead cells)] was used to document the degree of degeneration seen on histology. Immunohistochemistry was performed using a primary antibody to sPLA2 [(G-15):sc-14468], and the Goat Immunocruz Staining System sc-2053 (Santa Cruz Biotechnology, Inc.) was used to complete the staining. Jejunum was used as the positive control tissue to optimize the staining and a primary antibody dilution of 1: 20 was established as optimal. The percentage of cell staining was noted (1 < 50%, 2 >50%) and a semiquantitative three-point scoring system (0, little or none; 2, moderate; 3, strong) was used to document the intensity of staining seen. The assessment was performed by a consultant histopathologist (C.K.) who was unaware of the identity or clinical characteristics of the patients from whom the sample had been obtained.
Outcome was assessed at 2 months according to the Modified Stauffer–Coventry's evaluation criteria (Table 1).24 ‘Good’ and ‘excellent’ outcomes were considered to represent a satisfactory outcome, and ‘fair’ and ‘poor’ an unsatisfactory outcome.
A preoperative pain perception threshold of 6.1 mA (SD 3.1) has previously been demonstrated over the ipsilateral affected dermatome of patients scheduled for lumbar discectomy.27 A 50% lesser preoperative pain perception threshold (PPT) in those with an unsatisfactory outcome was assumed, resulting in a standardized effect size of 1. An unsatisfactory outcome was anticipated in 35% patients1,2 and a minimal sample size of 40 individuals was calculated. The Mann–Whitney U test, the Fisher's exact test, the chi-squared test, Wilcoxon signed rank test and Bland–Altman plots were used as appropriate and P < 0.05 was considered significant.
Forty-three patients were recruited. Two underwent a surgical procedure other than the planned lumbar discectomy and were excluded. Another two were not contactable following discharge from hospital and data from these patients were not analysed. Ten (26%) of the remaining 39 patients had an unsatisfactory outcome.
A total of 429 mean measurements of sensation and pain perception thresholds were available for analysis. For 322 (75.1%) of these, the first three consecutive measurements of the sensation and pain perception thresholds were satisfactory and the mean of these was calculated. One additional measurement was necessary in 93 (21.7%) instances, two additional measurements were taken on 13 (3%) occasions, and once (0.2%) three additional measurements were required in order to attain three nonconsecutive measurements such that the greater was no more than 120% of the least for either the sensation or pain perception threshold. One patient who had a satisfactory outcome refused all data collection 4 h postoperatively. A further six patients (five with a satisfactory outcome) declined the 4 h postoperative quantitative sensory testing and two patients (one with a satisfactory outcome) declined the 24 h postoperative quantitative sensory testing. Four patients (two with an unsatisfactory outcome) were unable to attend for assessment 2 months postoperatively, but they returned the HADS and McGills pain questionnaires by post.
Patients with satisfactory and unsatisfactory outcomes were similar with respect to patient characteristics (Table 2), duration of pain preoperatively, involvement in litigation as a result of the back pain, employment status, smoking status, disc level, side and site of pain, root tension signs and whether disc contents were found to be extruded at the time of surgery (data available on request).
Ten healthy volunteers (seven women, three men, mean age of 36.6 +/− 9.7 years) were recruited to assess inter-tester reliability of the quantitative sensory testing. On analysis of the Bland–Altman plots, all measurements of the sensation perception threshold (SPT) and the PPT lag were within 1.96 standard deviations of the mean, thus indicating an acceptable level of agreement between the two observers (data available on request).
Patients with a satisfactory outcome had a greater preoperative pain perception threshold to an electrical stimulus over the affected ipsilateral (P = 0.032) and affected contralateral dermatome (P = 0.021) than those with an unsatisfactory outcome (Table 2). The ratio of the pain perception threshold at the L4 or L5 dermatome to that at the C5 dermatome on the contralateral side [(L4 or L5 –C5)/C5×100] was greater preoperatively in patients with a satisfactory outcome than in those with an unsatisfactory outcome (mean rank 22.3 versus mean rank 13.3; Wilcoxon signed rank test, P = 0.031). Patients with satisfactory and unsatisfactory outcomes were similar in terms of the neurophysiological parameters measured at the other time points (data available on request).
In patients with a satisfactory outcome, the pain perception threshold decreased from preoperatively to 2 months postoperative at the ipsilateral (P = 0.004) and contralateral (P = 0.03) affected dermatomes (Table 3). In patients with a satisfactory outcome, the pain perception threshold at the ipsilateral affected dermatome decreased from preoperatively until 4 (mean rank 3.3 versus mean rank 2.2; Wilcoxon signed rank test, P = 0.008) and 18 h (mean rank 4.6 versus mean rank 2.4; Wilcoxon signed rank test, P = 0.003) postoperatively. No such changes were observed in patients with an unsatisfactory outcome.
Patients with and without a satisfactory outcome were similar with respect to nitric oxide metabolite concentrations at each time point (data available on request). However, plasma concentrations of nitric oxide metabolites decreased in patients with a satisfactory outcome from preoperatively to 18 h postoperatively (median 5.9, Wilcoxon signed rank test, P = 0.010), but not in those with an unsatisfactory outcome (median 9.2, Wilcoxon signed rank test, P = 0.074). A similar change was not seen from preoperatively to immediate postoperatively (satisfactory outcome: median 0.8, P = 0.9; unsatisfactory outcome: median 1.9, P = 0.51) or preoperatively to 4 h postoperatively (satisfactory outcome: median 2.9, P = 0.092; unsatisfactory outcome: median 4.4, P = 0.093).
Two months postoperatively, those who reported an unsatisfactory outcome achieved greater scores on the VAS for pain (P = 0.013) and the McGill Pain Questionnaire [pain rating index (PRI): P < 0.001; present pain index (PPI): P < 0.001, Table 4]. These patients also had greater scores on the VAS for anxiety (P = 0.039) and the HADS score (anxiety, depression and total scores, all P = 0.002, Table 4) 2 months postoperatively. Those with an unsatisfactory outcome had greater preoperative HAD scores (anxiety P = 0.017; total P = 0.044). They also had greater pain scores 4 h postoperatively (VAS pain P = 0.026) and 24 h postoperatively (VAS pain P = 0.030; McGill Pain Questionnaire: PRI P = 0.043; PPI P = 0.015, Table 4). A significant decrease in preoperative measures of anxiety, depression and pain were noted 2 months postoperatively in those with a satisfactory outcome (HAD score, VAS for pain, McGill's PRI and PPI, all P < 0.001; VAS anxiety P = 0.001, Table 3). Patients with and without a satisfactory outcome were similar with respect to analgesic requirements and adverse effects from analgesic consumption (data available on request).
The intensity and the percentage of cells staining for PLA2 and the degree of necrosis were similar in the two groups (data available on request). Staining was mostly seen intracellularly in nonnecrotic tissue; however, strong staining for PLA2 was seen in extracellular tissue in an area of marked necrosis from a patient who had a satisfactory outcome. Staining with endothelial markers CD31 and 34 was negative, indicating that the PLA2 detected was extravascular.
Ten (26%) of the patients had an unsatisfactory outcome 2 months after lumbar discectomy. Compared with those with an unsatisfactory outcome, patients with a satisfactory outcome demonstrated a greater pain perception threshold over the affected dermatome preoperatively. Unlike patients with an unsatisfactory outcome, patients with a satisfactory outcome demonstrated a decrease in the pain perception threshold at the affected dermatome and the serum nitric oxide metabolite concentrations during the perioperative period. An unsatisfactory outcome was associated with greater preoperative anxiety and depression and greater pain in the acute postoperative period. These finding are consistent with previous reports.28–33 We did not identify an association between the amount or intensity of immunostaining for PLA2 in disc tissue and patient outcome, consistent with previous findings.34–37
Neuroplastic changes can result in complex and apparently conflicting findings on quantitative sensory testing; for example, certain patients with neuropathic pain demonstrate a decrease in pain thresholds, which may represent central sensitization, whereas others display increased thresholds, possibly reflective of nociceptive deafferentation.38 Patients presenting for lumbar discectomy with leg pain only demonstrated greater pain perception thresholds to an electrical current than those with leg and back pain, and those with back pain only demonstrated the lowest thresholds. It was suggested that the different types of pain represented diverse central neuroplastic changes (including inhibition and facilitation).39 The initially greater and subsequent decrease in PPT observed in our patients with a satisfactory outcome may result from a reversal of deafferentation through surgical decompression. The absence of such changes in those with unsatisfactory outcomes may be owing to central sensitization, a process less likely to be reversed by surgery.
In this study, patients with a satisfactory outcome demonstrated a greater pain perception threshold preoperatively at the affected dermatome than those who had an unsatisfactory outcome. They also demonstrated a ratio of the pain perception threshold at the affected dermatome to that at the contralateral C5 dermatome. Although in our study and previously31 great interindividual variation exists in pain perception threshold, these findings indicate that referencing the pain perception threshold measured in the symptomatic dermatome to the contralateral C5 dermatome pain perception threshold may provide a means of predicting postoperative outcome. In addition, an unsatisfactory outcome was predicted by a greater preoperative HADS score (P = 0.044). It may be possible to devise a predictive index using preoperative factors to predict postoperative outcome. Further work is necessary to determine whether these findings can be replicated in a larger group of patients.
Better outcome following surgical treatment for lumbar disc degeneration (including lumbar disc herniation and lumbar spinal canal stenosis) correlates with lesser preoperative levels of nitric oxide metabolites in the cerebrospinal fluid (CSF), and with a decrease perioperatively in these levels.40 Also, nitric oxide metabolite levels in the CSF are higher in those with a longer duration of sciatica.41 Plasma concentrations of stable nitric oxide metabolites are greater in dysmenorrhoeal individuals than those without,42 suggesting a role for systemic nitric oxide in pain syndromes. We did not demonstrate an association between absolute values for preoperative serum nitric oxide metabolite levels and outcome. However, unlike those with an unsatisfactory outcome, we found that patients with a satisfactory outcome showed a decrease in serum nitric oxide metabolites from preoperative to 18 h postoperative levels. This may be attributable to an attenuation of an inflammatory process following surgery in those with a satisfactory outcome. It is possible that the surgical removal of herniated lumbar disc tissue may remove a source of nitric oxide production,13–15 with a resulting decrease in symptoms for some patients.
This study has certain limitations. The range of duration of preoperative pain was 4 weeks–10 years, indicating marked heterogeneity in the sample studied. One might interpret the differences observed as those robust enough to outweigh other confounding factors. Although four investigators participated in the collection of quantitative sensory testing data, all were trained and adhered to written guidelines. Good interobserver agreement between measurements taken by two investigators was demonstrated on analysis of observations. Also, as this is an exploratory trial, no correction was made for multiple comparisons.
Of patients undergoing lumbar discectomy for low-back pain, those with an unsatisfactory outcome experience greater preoperative anxiety and greater pain during the early postoperative period. Those with a satisfactory outcome demonstrate greater preoperative pain perception thresholds at the affected dermatome unlike those with an unsatisfactory outcome. Patients with a satisfactory outcome demonstrate a decrease in pain perception thresholds and plasma concentration of stable nitric oxide metabolites during the perioperative period. These findings justify a larger prospective observational study, the objective of which is the identification and validation of a multiparameter tool for predicting outcome (and patient satisfaction) from lumbar discectomy.
The authors would like to thank Dr Kathleen O'Sullivan, Director of the Statistical Consultancy Unit, School of Mathematical Sciences, University College Cork, for her substantial contribution.
Support was provided solely by institutional sources.
There are no conflicts of interest.
Financial Support given by the Department of Anaesthesia and University College Cork, Cork City, Ireland.
1 Deyo RA, Tsui-Wu YJ. Descriptive epidemiology of low-back
pain and its related medical care in the United States. Spine 1987; 12:264–268.
2 Gibson JN, Grant IC, Waddell G. The Cochrane review of surgery for lumbar disc prolapse and degenerative lumbar spondylosis. Spine 1999; 24:1820–1832.
3 Atlas SJ, Deyo RA, Keller RB, Chapin AM, et al.
The Maine Lumbar Spine Study, part II. 1-year outcomes of surgical and nonsurgical management of sciatica. Spine
4 Junge A, Fronlich M, Ahrens S, et al
. Predictors of bad and good outcome of lumbar spine surgery. A prospective clinical study with 2 years follow-up. Spine 1996; 21:1056–1065.
5 Asch HL, Lewis PJ, Moreland DB, et al
. Prospective multiple outcomes study of outpatient lumbar discectomy: should 75 to 80% success rates be the norm? J Neurosurg 2002; 96(1 Suppl):34–44.
6 Hakkinen A, Ylinen J, Kautiainen H, et al
. Does the outcome 2 months after lumbar disc surgery predict the outcome 12 months later? Disabil Rehabil 2003; 25:968–972.
7 Freynhagen R, Baron R, Tolle T, et al
. Screening of neuropathic pain components in patients with chronic back
pain associated with nerve compression: a prospective observational pilot study (MIPORT). Curr Med Res Opin 2006; 22:529–537.
8 Flor H, Braun C, Elbert T, Birbaumer N. Extensive reorganization of primary somatosensory cortex in chronic back
pain patients. Neurosci Lett 1997; 224:5–8.
9 Nygaard OP, Mellgren SI. The function of sensory nerve fibres in lumbar radiculopathy. Spine 1998; 23:348–353.
10 Yamashita T, Kanaya K, Sekine M, et al
. A quantitative analysis of sensory function in lumbar radiculopathy using current perception threshold testing. Spine 2002; 27:1567–1570.
11 Ngaard OP, Kloster R, Solberg T, Mellgren SI. Recovery of function in adjacent nerve roots after surgery for lumbar disc herniation: use of quantitative sensory testing in the exploration of different populations of nerve fibres. J Spinal Disord 2000; 13:427–431.
12 Nygaard OP, Kloster R, Mellgren SI. Recovery of sensory nerve fibres after surgical decompression in lumbar radiculopathy: use of quantitative sensory testing in the exploration of different populations of nerve fibres. J Neurol Neurosurg Psychiatry 1998; 64:120–123.
13 Kang JD, Georgescu HI, McIntyre-Larkin L, et al
. Herniated lumbar intervertebral discs spontaneously produce matrix metalloproteinases, nitric oxide
, interleukin-6 and prostaglandin E2. Spine 1996; 21:271–277.
14 Kang JD, Stefanovic-Racic M, McIntyre LA, et al
. Towards a biochemical understanding of human intervertrebral disc degeneration and herniation. Spine 1997; 22:1065–1073.
15 Hashizume H, Kawakami M, Nishi H, Tamaki T. Histochemical demonstration of nitric oxide
in herniated lumbar discs. A clinical and animal model study. Spine 1997; 22:1080–1084.
16 Saal JS, Franson RC, Dobrow R, et al
. High level of inflammatory phospholipase A2 activity in lumbar disc herniations. Spine 1990; 15:674–678.
17 Franson RC, Saal JS, Saal JA. Human disc phospholipase is inflammatory. Spine 1992; 17:S129–S132.
18 Ozaktay AC, Cavanaugh JM, Blagoev DC, King AI. Phospholipase A2-induced electrophysiological and histological changes in rabbit dorsal lumbar spine tissues. Spine 1995; 20:2659–2668.
19 Kawakami M, Tamaki T, Weinstein JN, et al
. Pathomechanism of pain-related behaviour produced by allografts of intervertebral disc in the rat. Spine 1996; 21:2101–2107.
20 Kawakami M, Tamaki T, Hashizume H, et al
. The role of Phospholipase A2 and nitric oxide
in pain related behaviour produced by an allograft of intervertebral disc material to the sciatic nerve of the rat. Spine 1997; 22:1074–1079.
21 Chen C, Cavanaugh JM, Ozaktay AC, et al
. Effects of phospholipase A2 on lumbar nerve root structure and function. Spine 1997; 22:1057–1064.
22 Sawin PD, Traynelis VC, Rich G, et al
. Chymopapain-induced reduction of proinflammatory phospholipase A2 activity and amelioation of neuropathic behavioral changes in an in vivo model of acute sciatica. J Neurosurg 1997; 86:998–1006.
23 Ozaktay AC, Kallakuri S, Cavanaugh JM. Plospholipase A2 sensitivity of the dorsal root and dorsal root ganglion. Spine 1998; 23:1297–1306.
24 Stauffer RN, Coventry MB. Anterior interbody lumbar fusion: analysis of mayo clinic series. J Bone Joint Surg Am 1972; 54:756–768.
25 Yang BK, Vivas EX, Reiter CD, Gladwin MT. Methodologies for the sensitive and specific measurement of S-nitrosothiols, iron-nitrosyls, and nitrite in biological samples. Free Radic Res 2003; 37:1–10.
26 Melzack R. The short-form McGill pain questionnaire. Pain 1987; 30:191–197.
27 Wilder-Smith OHG, Tassonyi E, Senly C, et al
. Surgical pain is followed not only by spinal sensitization but also by supraspinal antinociception. Br J Anaesth 1996; 76:816–821.
28 Junge A, Dvorak J, Ahrens S. Predictors of bad and good outcomes of lumbar disc surgery. A prospective clinical study with recommendations for screening to avoid bad outcomes. Spine 1995; 20:460–468.
29 Schade V, Semmer N, Main CJ, et al
. The impact of clinical, morphological, psychosocial and work-related factors on the outcome of lumbar discectomy. Pain 1999; 80:239–249.
30 Sorensen LV, Mors O, Skovlund O. A prospective study of the importance of psychological and social factors for the outcome after surgery in patients with slipped lumbar disc operated upon for the first time. Acta Neurochir 1987; 88:119–125.
31 Graver V, Ljunggren AE, Malt UF, et al
. Can psychological traits predict the outcome of lumbar disc surgery when anamnestic and physiological risk factors are controlled for? Results of a prospective cohort study. J Psychosom Res 1995; 39:465–476.
32 Kjellby-Wendt G, Styf JR, Carlsson SG. The predictive value of psychometric analysis in patients treated by extirpation of lumbar intervertebral disc herniation. J Spinal Disord 1999; 12:375–379.
33 Edwards RR, Klick B, Buenaver L, et al
. Symptoms of distress as prospective predictors of pain-related sciatica treatment outcomes. Pain 2007; 130:47–55.
34 Piperno M, le Graverand MPH, Reboul P, et al
. Phospholipase A2 activity in herniated lumbar discs. Clinical correlations and inhibition by piroxicam. Spine 1997; 22:2061–2065.
35 Miyahara K, Ishida T, Hukuda S, et al
. Human group II phospholipase A2 in normal and diseased intervertebral discs. Biochim Biophys Acta 1996; 1316:183–190.
36 Gronblad M, Virri J, Ronkko S, et al
. A controlled biochemical and immunohistochemical study of human synovial-type (group II) phospholipase A2 and inflammatory cells in macroscopically normal, degenerated, and herniated human lumbar disc tissues. Spine 1996; 21:2531–2538.
37 Habtemariam A, Gronblad M, Virri J, et al
. Comparative immunohistochemical study of group II (synovial type) and group IV (cytosolic) phospholipase A2 in disc prolapse tissue. Eur Spine J 1998; 7:387–393.
38 Baumgartner U, Magerl W, Klein T, et al
. Neurogenic hyperalgesia versus painful hypoalgesia: two distinct mechanisms of neuropathic pain. Pain 2002; 96:141–151.
39 Wilder-Smith OHG, Tassonyi E, Arendt-Nielsen L. Preoperative back
pain is associated with diverse manifestations of central neuroplasticity. Pain 2002; 97:189–194.
40 Kimura S, Watanabe K, Yajiri Y, et al
. Cerebrospinal fluid nitric oxide
metabolites are novel predictors of pain relief in degenerative lumbar diseases. Pain 2001; 92:363–371.
41 Kimura S, Watanabe K, Yajiri Y, et al
. Cerebrospinal fluid nitric oxide
metabolites in painful diseases. Neuroreport 1999; 10:275–279.
42 Sun MF, Huang HC, Lin SC, et al
. Evaluation of nitric oxide
and homocysteine levels in primary dysmenorrheal women in Taiwan. Life Sci 2005; 76:2005–2009.