Carnosine Treatment for Gulf War Illness: A Randomized Controlled Trial

About 25% of 1990-1991 Persian Gulf War veterans experience disabling fatigue, widespread pain, and cognitive dysfunction termed Gulf War illness (GWI) or Chronic Multisymptom Illness (CMI). A leading theory proposes that wartime exposures initiated prolonged production of reactive oxygen species (ROS) and central nervous system injury. The endogenous antioxidant L-carnosine (β-alanyl-L-histidine) is a potential treatment since it is a free radical scavenger in nervous tissue. To determine if nutritional supplementation with L-carnosine would significantly improve pain, cognition and fatigue in GWI, a randomized double blind placebo controlled 12 week dose escalation study involving 25 GWI subjects was employed. L-carnosine was given as 500, 1000, and 1500 mg increasing at 4 week intervals. Outcomes included subjective fatigue, pain and psychosocial questionnaires, and instantaneous fatigue and activity levels recorded by ActiWatch Score devices. Cognitive function was evaluated by WAIS-R digit symbol substitution test. Carnosine had 2 potentially beneficial effects: WAIS-R scores increased significantly, and there was a decrease in diarrhea associated with irritable bowel syndrome. No other significant incremental changes were found. Therefore, 12 weeks of carnosine (1500 mg) may have beneficial cognitive effects in GWI. Fatigue, pain, hyperalgesia, activity and other outcomes were resistant to treatment.

related dipeptide that may scavenge ROS in the central nervous system (Boldyrev, Dupin, Bunin, Babizhaey, & Severin, 1987;Kohen, Yamamoto, Cundy, & Ames, 1988) and act as an antioxidant and buffering ion in muscle and possibly other peripheral tissues (Abe, 2000;Holliday & McFarland, 2000). Carnosine may promote brain health because autistic children taking 800 mg a day demonstrated significant cognitive improvements (Chez et al., 2002). This suggested that carnosine may cross the blood brain barrier.

Study Design
This randomized double blind placebo controlled pilot study of carnosine (500 mg capsules, Village Green Apothecary, Bethesda, MD; Food and Drug Administration Investigational New Drug FDA IND # 101,700) versus indistinguishable microcrystalline cellulose capsules was tested as add-on therapy in GWI. Dose escalation started with 500 mg once daily, and increased to 500 mg twice daily, and then 1,000 mg in the morning and 500 mg in the evening (1,500 mg/day) at 4 week intervals (ClinicalTrials.gov identifier NCT00810368). Sets of 4 x 100 capsule bottles were prepared and blocks with 3 placebo and 3 carnosine sets randomly mixed. Each set was assigned a sequential 6 digit alpha-numerical identification code so study drug could be dispensed by research pharmacists not involved in blinding the study. After the week 12 visit, the 3 capsules per day were tapered by 1 per week until discontinued. The study was approved by the Georgetown University Institutional Review Board (ID # 2008-068). Data analysis, recruitment, protocol implementation, and patient participation all occurred at Georgetown University Medical Center and Hospital in Washington, D.C. of the United States.

Subject Recruitment
Subjects were recruited by on-line and newspaper advertisements, and other contacts. Telephone informed consent was obtained to determine eligibility and contact information. GWI subjects were eligible if they met the 1998 CDC criteria (Fukuda et al., 1998;RAC-GWVI, 2008) or had been diagnosed with chronic fatigue syndrome after Persian Gulf deployment (Gray, Reed, Kaiser, Smith, & Gastanaga, 2002). Exclusions included active military personnel, pregnant women, children, persons suffering from major chronic illnesses or psychiatric disorders, and persons with metal implants or other exclusions from fMRI testing (Fukuda et al., 1998;Kang et al., 2009;RAC-GWVI, 2008;Spencer et al., 1998;McCauley et al., 2002;Steele, 2000).

Study Questionnaires
The CFS Severity score allowed subjects to score the cardinal features of fatigue, cognition, sore throat, lymph nodes, myalgia, arthralgia, sleep and exertional exhaustion (Fukuda et al., 1994;Reeves et al., 2003;Reeves et al., 2005) as none (0), trivial (1), mild (2), moderate (3), or severe (4) (Baraniuk et al., 1998;Baraniuk et al., 2000a;Naranch et al., 2002;Wasserfallen, Gold, Schulman, & Baraniuk, 1999) . The maximum sum for this severity score was 32. Medical Outcome Survey Short Form 36 (SF-36) domains were assayed for changes in quality of life (Ware & Sherbourne, 1995). Additional secondary symptom and psychometric questionnaires were completed to corroborate and expand data on GWI symptom co-variance and overlap with other syndromes (see Supplementary online material, SOM). The instruments were assessed for magnitudes, standard deviations, and reproducibility between weeks 0 and 12 to determine which may be of value for future studies. Subjects could complete the questionnaires on paper or our on-line, Google-based system. Each subject received a personal log-in, password and confidential identification code to access and score the questionnaires (Zheng et al., 2011a;Zheng et al., 2011b). No personal identifying information was collected over the internet. All data were stored in confidential fashion on a password protected computer.
Subjects had physical examination, counting of manual tender points (Wolfe et al., 1990) and dolorimetry (Naranch et al., 2002). Screening blood work was drawn for HIV, pregnancy, blood counts, transaminases, and markers of inflammation. Cognitive function was assessed by Trail Making tests A and B (Corrigan & Hinkeldey, 1987) and digit symbol substitution test (Joy, Fein, Kaplan, & Freddman, 2000). Subjects toured the Center for Function and Molecular Imaging (CFMI) and practiced 0-and 2-back testing until they felt competent. They returned later for their fMRI with cognitive testing.
Subjects wore ActiWatch-Score wrist devices (Gironda et al., 2007) for the first 2 weeks of the study, and weeks 12 to 14. In preliminary tests, it was difficult to discriminate between periods of rest and sleep related to chronic fatigue and sedentary work (e.g. computer) performed by control subjects. Diaries were consulted to confirm periods of rest and sleep. Average accelerometer values for each 1 hour epoch during sedentary periods were < 40 units (mean + 2 standard deviations). This threshold distinguished "rest" from "active" periods. Records were reviewed for times when the watches were removed (zero activity). The percent of active epochs (% Active Time) during the entire time the devices were worn was used as an outcome variable. The average activity levels during active periods were also calculated on the assumption that treatment may be associated with more vigorous or prolonged movement. The Score alarm buzzed randomly at approximately 1 hour intervals to remind subjects to grade their instantaneous fatigue on a scale of 0 (no fatigue) to 10 (worst fatigue in life) (Kim et al., 2001). The average fatigue score for each day and the standard deviations and coefficients of variation were calculated to assess the magnitude of daily fluctuations in this symptom, and the trends for fatigue within each 2 week period.
Contact was maintained during the study by telephone and e-mail. Subjects who developed adverse events or discontinued participation were quizzed for their rationale. Symptoms, laboratories, and questionnaires were assessed at the week 6 safety visit. The week 12 visit followed the same format as week 0.

Subjects and Adverse Events
Between February 2008 and November 2011 a total of 358 subjects were contacted about participation. Two hundred were interested and began the verbal screening process. Thirty-four subjects were enrolled and randomized. Nine subjects did not complete the study. Two were terminated. One developed an exacerbation of their typical severe musculoskeletal costochondritic chest pain, but was referred by their primary care physician to a cardiologist (Grade 1 adverse event, not related). Another was found to have elevated globulins at week 6. This participant did not return for follow-up or respond to any inquiries, and was lost to contact (Grade 2 adverse event, not related). Three subjects stopped participation within the first 2 weeks of the study and did not reply to any inquiries.
Three subjects did not want to continue in the study after they reached the 6 week mark because they were frustrated by the absence of any improvements. One moved away after 8 weeks. One subject was found to have an asymptomatic elevation of alanine-serine transaminase plus interval increase in CRP (remained in normal range). As per protocol, study drug was discontinued for 2 weeks. Hepatitis A, B and C serologies were negative, and tests normalized. Study drug was restarted and no alterations in symptoms or blood tests developed. This www.ccsen was consid subjects, w between th Geake & Hansen, 2010). In contrast, there were no changes for Trail Making tests A and B that may assess inter-hemispheric cognitive processing (Corrigan & Hinkeldey, 1987). If so, carnosine may have selective trophic or protective effects in discrete brain regions.
Carnosine treatment was also associated with reduced stool frequency indicating an improvement in diarrhea. Other questionnaire outcomes including the primary aims were not altered by either carnosine or placebo treatments. Subjects' levels of complaints were remarkably reproducible between week 0 and week 12. The symptoms caused significant disability based on SF-36 quality-of-life scores.
The absence of a placebo effect was important because the expectations of improvement did not translate into symptomatic improvement. This steadfast reporting demonstrated that GWI veterans were a highly reliable group who report their symptoms in a reproducible fashion. As a result, beneficial treatment effects in other studies are likely to be true results. The study drug was add-on therapy. The chronic severity of complaints suggests that current treatments had limited therapeutic benefits.
An extensive data set (SOM) was collected for (a) comparisons to CFS subjects, (b) potential characterization of CMI subtypes, (c) post hoc hypothesis generation, (d) selection of appropriate questionnaires for future studies, and (e) sample size calculations. Questionnaires assessed fatigue, pain, and interoceptive complaints such as rhinitis, dyspnea, irritable bladder and irritable bowel syndromes, and perceived responses to irritant exposures.
The cognitive improvement suggested by this study may be analogous to changes in autism. Carnosine given as 800 mg/d for 8 weeks in double blind, placebo controlled fashion to autistic children significantly improved behavior, socialization, communication and total scores on the Gilliam Autism Rating Scale (Chez et al., 2002).
Carnosine improves athletic capabilities, probably due to antioxidant effects or buffering capacity in muscle (Abe, 2000;Holliday & McFarland, 2000). Physiological stressors acutely increased avian muscle carnosine by 10-fold, but had no effect on brain levels (Manhiani et al., 2011). Similar stressor effects have not been studied in man.
Carnosine digestion by CNDP1 and CNDP2 releases β-alanine and histidine. The latter can chelate metals and scavenge singlet oxygen during ischemia (Abe, 2000;Holliday & McFarland, 2000;Lee et al., 1999). Increased histidine absorption into the brain was hypothesized to trigger an increase in brain homocarnosine synthesis by the rate limiting enzyme ATP-grasp domain-containing protein 1 (ATPGD1) in astroglial cells (Drozak, Veiga-da-Cunha, Vertommen, Stroobant, & Van Schaftingen, 2010). Lumbar punctures and measurements of histidine, carnosine and homocarnosine in cerebrospinal fluid would be required to confirm this hypothesis.
β-Alanine is the rate limiting substrate for carnosine synthesis by ATPGD1 in human and equine muscle, since histidine supplements do not boost carnosine levels (Dunnett & Harris, 1999). β-Alanine supplementation improved performance in anaerobic muscle testing (Hoffman et al., 2008). β-Alanine may delay the onset of neuromuscular fatigue in intensively exercising athletes (Artioli, Gualano, Smith, Stout, & Lancha 2010). Similar effects may not have been evident in our GWI subjects since they were not enrolled in structured aerobic or anaerobic training programs or tested for aerobic exercise capacity as part of this pilot study. High doses of β-alanine (3 and 6 g/d) improved oxygen utilization during high intensity exercise training in elite athletes (Smith et al., 2009;Walter, Smith, Kendall, Stout, & Cramer, 2010), although lower doses do not have effects on VO 2MAX during aerobic testing.
Oral doses greater than 800 mg of β-alanine may induce paresthesias (Artioli et al., 2010). This is of relevance to our study since β-alanine (Shinohara et al., 2004) and ATP may bind Mrgprd receptors of murine dorsal root ganglion neurons that mediate cutaneous mechanical nociception (Cavanaugh et al., 2009). We found a higher drop-out rate in carnosine than placebo treated GWI veterans. They did not report any changes in cutaneous perceptions when questioned, but a subliminal effect cannot be ruled out in those subjects who were lost to contact. Perceptions of enteric interoceptive sensations may have developed based on changes in gastrointestinal www.ccsenet.org/gjhs Global Journal of Health Science Vol. 5, No. 3; complaints. At week 0, half of the carnosine group reported irritable bowel syndrome with mixed diarrhea and constipation, but none had predominantly constipation. However, after 12 weeks constipation was present in 29%, and the rate of mixed diarrhea and constipation had dropped to 14%. The placebo group had no changes in irritable bowel complaints. Although it is important not to over interpret these findings in light of the small sample size, any treatment that would improve the severe gastrointestinal complaints of GWI would be welcomed.
The study was limited by the relatively small sample size. The putative treatment effects require validation in similar randomized double blind placebo-controlled studies with larger groups of participants. The small sample size was hampered even further by the nine subjects who withdrew. They had no apparent symptomatic changes or adverse events. This includes parasthesias that have been noted previously (Artioli et al., 2010).
Study compliance was excellent based on diaries and pill counts. However, there were no cerebrospinal fluid, serum, or urine measurements of carnosine, beta-alanine or histidine to confirm absorption, pharmacodynamics, or brain metabolism to homocarnosine. Elevations in brain GABA levels in the treatment group may be identified by molecular spectroscopy (analysis in progress).
Another limitation was the subjective nature of many of the outcome measures. This was by design for this pilot study to assess a broad set of symptomatic and psychometric variables that may have shown beneficial responses to this antioxidant therapy. For instance, even though we observed a possible significant improvement in symptoms of diarrhea, there are several different sets of criteria for IBS, and this treatment effect could be better confirmed with more precise measures for stool form and frequency. In addition, subjects' scores for many questionnaire items were at the upper limits, suggesting possible ceiling effects. Small improvements in their conditions may not have been detectable with the scales employed.
Hyperalgesia and disability leading to inactivity were assessed in semi-objective fashion by dolorimetry and actigraphy, respectively. These outcomes were insensitive to change. It is possible that 12 weeks of add-on therapy without adjunctive treatments was insufficient to reverse the consequences of long duration neural dysfunction. In particular, the chronicity of the subjects' complaints may have made it difficult for them to assess and identify subjective improvements. Established neural patterns may be resistant to modification. Improved understanding of the molecular pathophysiology is required in order to optimally select fMRI and other objective outcome measures for treatment studies such as this one.

Conclusion
This randomized double blind placebo controlled 12 week dose escalation trial of carnosine showed an increase in digit symbol substitution score and potentially decreased diarrhea in GWI. Similar larger studies with objectively defined and pathophysiologically based outcome measures will be required to confirm and extend these findings.

Author Contributions
J.N.B. designed the protocol. R.U.R, C.R.T. and J.NB recruited participants, administered and conducted the study. R.C., S.E.A, and J.N.B. interpreted and completed clinical data analysis. R.C, S.E.A., and J.N.B. wrote the manuscript. All authors edited the manuscript. None of the authors have financial or non-financial competing interests.