Journal of Molecular Medicine

, Volume 88, Issue 1, pp 85–92

Phase II nonrandomized study of the efficacy and safety of COX-2 inhibitor celecoxib on patients with cancer cachexia


    • Department of Medical OncologyUniversity of Cagliari
    • Cattedra di Oncologia MedicaUniversità di Cagliari, Azienda Ospedaliero-Universitaria di Cagliari
  • Antonio Macciò
    • Department of Medical OncologyUniversity of Cagliari
  • Clelia Madeddu
    • Department of Medical OncologyUniversity of Cagliari
  • Roberto Serpe
    • Department of Medical OncologyUniversity of Cagliari
  • Giorgia Antoni
    • Department of Medical OncologyUniversity of Cagliari
  • Elena Massa
    • Department of Medical OncologyUniversity of Cagliari
  • Mariele Dessì
    • Department of Medical OncologyUniversity of Cagliari
  • Filomena Panzone
    • Department of Medical OncologyUniversity of Cagliari
Original Article

DOI: 10.1007/s00109-009-0547-z

Cite this article as:
Mantovani, G., Macciò, A., Madeddu, C. et al. J Mol Med (2010) 88: 85. doi:10.1007/s00109-009-0547-z


Chronic inflammation is one of the main features of cancer cachexia. Experimental and clinical studies showed that cyclooxygenase-2 inhibitors, such as celecoxib, may be beneficial in counteracting major symptoms of this devastating syndrome. We carried out a prospective phase II clinical trial to test the safety and effectiveness of an intervention with the COX-2 inhibitor celecoxib (300 mg/day for 4 months) on key variables of cachexia (lean body mass, resting energy expenditure, serum levels of proinflammatory cytokines, and fatigue) in patients with advanced cancer at different sites. A sample of 24 patients was enrolled from January to December 2008 and all were deemed assessable. A significant increase of lean body mass and a significant decrease of TNF-α were observed. Moreover, an improvement of grip strength, quality of life, performance status, and Glasgow prognostic score was shown. There were no grade 3/4 toxicities. Patient compliance was very good; no patient had to reduce the celecoxib dosage nor interrupt treatment. Our results showed that the COX-2 selective inhibitor celecoxib is an effective single agent for the treatment of cancer cachexia. Although the treatment of cancer cachexia, a multifactorial syndrome, is more likely to yield success with a multitargeted approach; in the present study, we were able to show that a treatment, such as celecoxib, addressing a single target, albeit very important as chronic inflammation, could have positive effects. Therefore, phase III clinical trials are warranted to test the efficacy and safety of celecoxib.


Cancer cachexiaChronic inflammationCOX-2 inhibitorsCelecoxibLean body mass, resting energy expenditureProinflammatory cytokinesFatigue


Cachexia is a multifactorial syndrome characterized by tissue wasting, loss of body weight, substantially due to loss of lean body mass (LBM), metabolic alterations, fatigue, reduced performance status [1], very often accompanied by anorexia, leading to a reduced food intake. It accompanies the end stage of several chronic diseases, especially cancer, and therefore, it is termed “cancer-related anorexia/cachexia syndrome” (CACS) [25]. At the time of cancer diagnosis, 80% of patients with upper gastrointestinal cancers and 60% with lung cancer have already had substantial weight loss. The prevalence of cachexia increases from 50% to 80% before death, and cachexia is the cause of death in 20% of the patients [6].

CACS is distinct from starvation in that it involves preferential wasting of LBM, while visceral proteins and fat mass are relatively preserved: its hallmarks are the features of systemic chronic inflammation, including increased production/release of acute-phase proteins and inflammatory cytokines, as well as tumor-derived catabolic factors [7]. Proinflammatory cytokines interleukin (IL)-1, IL-6, and tumor necrosis factor (TNF)-α play a central role in the pathophysiology of CACS [813]: there is evidence that a chronic, low-grade, tumor-induced activation of the host immune system, which shares several characteristics with the ‘‘acute-phase response’’ found after major traumatic events and septic shock, is involved in CACS [14]. Therefore, anti-inflammatory drugs such as cyclooxygenase (COX)-2 inhibitors could break the “vicious circle” leading to the onset and worsening of this devastating syndrome [15].

COX-2 is a bifunctional enzyme that has both cyclooxygenase and peroxidase activities, the former catalyzing the synthesis of prostaglandins from arachidonic acid. Selective COX-2 inhibitors inhibit prostaglandin biosynthesis (anti-COX-2 activity) but do not, or only partially, affect the peroxidase activity of COX, which can generate proximate carcinogens. Significant preclinical evidence strongly supports the potential role for these inhibitors in the treatment of cancer. Currently, COX-2 inhibitors are being studied in clinical trials to confirm their role in the prevention of cancer, particularly, colon cancer and in combination with chemotherapy and radiation therapy, to prove their effectiveness in cervical cancer, lung cancer, and brain tumors [16].

Increasing evidence suggests that eicosanoids such as prostaglandin (PG)E2 are involved in anorexia and cachexia development [17] and that COX-2 inhibitors could have a potential role in counteracting cancer cachexia [18]. Review of the literature showed studies from as early as 1975 demonstrating that indomethacin, a nonsteroidal anti-inflammatory drug (NSAID), was able to inhibit levels of urinary PGE2 metabolites and reduce serum hypercalcemia associated with cachexia [19]. Prostaglandin release is a component of the signaling cascade in skeletal muscle protein turnover in vitro, suggesting that COX-2 may play a role in pathologies of muscle catabolism and therefore, wasting [1]. Besides direct effects on muscle biology, prostaglandins can also regulate the expression of proinflammatory cytokines, such as IL-6 and TNF-α. NSAIDs have been shown to block catabolic effects on protein metabolism by serum from cachectic mice [20].

In an experimental model in mice, Hussey and Tisdale [21], studying the effect of the COX-2 inhibitor meloxicam, demonstrated that it was able to effectively attenuate cachexia in the murine adenocarcinoma (MAC)-16 possibly by exerting a direct effect on skeletal muscle protein degradation. In the tumor model mouse colon-26, both IL-6 and parathyroid hormone-related protein (PTHrP) induced by COX-2 upregulation were shown to be involved in the development of cachexia, and a COX-2 inhibitor (NS-398) was shown to be able to suppress the PTHrP production [22]. In mice bearing LP07 lung adenocarcinoma developing paraneoplastic syndromes such as cachexia, leukocytosis, and hypercalcemia, the combination of a NSAID, such as indomethacin and medroxyprogesterone acetate, showed to be effective in diminishing paraneoplastic symptoms and systemic inflammatory response including cachexia [23].

Based both on the above reported experimental findings and on clinical evidence (see under “Discussion” section), our hypothesis was that the administration of COX-2 inhibitor celecoxib to patients with cancer cachexia could improve body composition and function and mediate an improvement of patient quality of life (QL) mainly by down-regulating the mediators of chronic systemic inflammation.

We report the results of a prospective, phase II clinical trial of the efficacy and safety of the COX-2 inhibitor celecoxib on cachectic patients with advanced cancer at different sites. Primary efficacy endpoints were LBM, resting energy expenditure [24], serum levels of proinflammatory cytokines, and fatigue.

Patients and methods

Study design

The study was a phase II nonrandomized open label clinical trial with the aim of assessing the efficacy and safety of a 4-month intervention with the oral selective COX-2 inhibitor celecoxib in improving “key” variables (primary endpoints) of CACS: increase of LBM, decrease of resting energy expenditure (REE), decrease of IL-6 and TNF-α, and decrease of fatigue. The protocol was approved by the reference Ethics Committee. Written informed consent was obtained from all patients. The procedures followed were in accordance with good clinical practices and the Helsinki Declaration.

Eligibility and exclusion criteria

Patients with advanced cancer at different sites referring to the Department of Medical Oncology at the University of Cagliari, Italy were enrolled. Patient eligibility criteria were age range of 18–80 years, histologically confirmed tumor of any site at an advanced stage, loss of ≥5% of the ideal (or preillness) body weight in the previous 3–6 months and abnormal values of proinflammatory cytokines, and a life expectancy of ≥4 months. Patients could be receiving concomitant antineoplastic chemotherapy or hormone therapy with palliative intent or supportive care. Exclusion criteria were women of child-bearing age, positive history of heart disease, i.e., severe (New York Heart Association Functional class III or IV) heart failure, or known ejection fraction ≤35%, uncontrolled hypertension (systolic blood pressure >140 mmHg and diastolic blood pressure >90 mmHg), history of previous myocardial infarction, unstable angina, coronary revascularization, uncontrolled arrhythmia, congestive heart failure and cerebrovascular accident, previous gastrointestinal inflammatory disease and history of gastrointestinal ulceration, mechanical obstruction to feeding, and medical treatments inducing significant changes of patient metabolism or body weight.

Treatment plan

All patients included in the study were given celecoxib (Celebrex, Pfizer Italia, Latina, Italy) 300 mg/day, i.e., one 200 mg capsule plus one 100 mg capsule per day, taken orally for 4 months. Compliance was determined by study personnel counting the remaining pills at the end of each month for each patient. The dosage of 300 mg/day was chosen on the basis of the following rationale. Our phase II study [25], which tested the efficacy against cachexia of a combination approach including celecoxib at 200 mg/day showed positive results. The other two similar studies by Cerchietti et al. [26] and Couch et al. [27] tested the administration of 400 mg/day. None of these studies showed any grade of toxicity attributable to celecoxib. However, possible safety concerns raised by FDA in the past few years ( have led us to choose the more cautious dosage of 300 mg/day.

Efficacy endpoints

The following endpoints were assessed before treatment and at 1, 2, and 4 months after treatment start.

Primary efficacy endpoints

LBM was assessed both by bioelectrical impedance analysis (BIA, Bioelectric Impedance Analyser 101, Akern, Florence, Italy) and dual energy X-ray absorptiometry (DEXA, Hologic Delphi W scanner, Hologic Inc, Bedford, Massachusets, USA); the latter is currently considered the most reliable method. The REE was assessed by indirect calorimetry (Medgem, SensorMedics Italia Srl, Milan, Italy), which measures oxygen consumption per time unit. Serum levels of proinflammatory cytokines IL-6 and TNF-α were measured by enzyme-linked immunosorbent assays (Immunotech, Marseille, France). Fatigue was assessed by the multidimensional fatigue symptom inventory-short form (MFSI-SF) and calculated by a numerical score (total fatigue scores ranging from −24 to +96) [28, 29]; results are reported as mean scores.

Secondary efficacy endpoints

The following secondary efficacy endpoints were assessed: body weight and BMI; appetite by visual analog scale; grip strength by dynamometer (Jamar Hydraulic Hand Dynamometer, Sammons Preston, Bolingbrook, Illinois, USA); blood levels of reactive oxygen species (FORT test, Callegari SpA, Italy) and antioxidant enzyme glutathione peroxidase (Randox, Crumlin, UK) as markers of oxidative stress often associated with CACS [25, 30] by photometer; quality of life by the EORTC-QLQ-C30, EuroQL-5D (EQ-5Dindex and EQ-5DVAS); performance status [1] according to the Eastern Cooperative Oncology Group (ECOG) PS scale [24]; Glasgow prognostic score (GPS), currently considered a significant predictive index for survival in advanced stage cancer patients such as those cachectic [31, 32]; objective clinical response at the end of treatment (complete response, partial response, stable disease, and progressive disease) according to response evaluation criteria in solid tumors (RECIST) [33]; progression-free survival from the date of treatment start to the first evidence of progressive disease; and overall survival from the date of treatment start to the patient death or last visit at follow up.

Safety endpoints

The safety endpoints were classified as adverse events according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events v3.0 criteria [34]. As for cardiac monitoring, patients underwent an accurate cardiac evaluation including clinical visit, electrocardiogram, and echocardiography assessment at baseline and monthly thereafter.

Statistical analysis

The statistical objectives of the study were to analyze for statistically significant differences; the changes (the mean preintervention versus postintervention values) in treated patients for each primary and secondary efficacy endpoint.

One way analysis of variance was used for all comparisons, and the Wilcoxon sum rank test was used where appropriate. Statistical significance was established at p < 0.05. SPSS version 15.0 was used.



A sample of 24 patients was enrolled from January to December 2008; all were deemed assessable. Patient clinical characteristics are listed in Table 1. The men/women ratio was well balanced (13/11). The mean age was 60.6 years. The most frequent tumor sites were head and neck (six patients), lung (six patients), and colorectal (six patients). Patients were >90% stage IV and GPS was high (two) in 50% of them. Six out of 24 patients (25%) had a weight loss, although, less than 5% of them in the previous 3 months. Therefore, they were not strictly cachectic; however, they had very high levels of proinflamamtory cytokines (IL-6 and TNF-α), which, according to our experience and almost general consensus, are deemed to be a very significant variable predictive of the onset of cachexia.
Table 1

Patient clinical characteristics










Age (years); mean ± SD



Weight (kg); mean ± SD

61 ± 11.8














Weight loss

<5% (3–6 months)



5–10% (3–6 months)



>10% (3–6 months)



Tumor site

Head and neck






































Glasgow prognostic score




1-albumine <32 g/l



1-CRP >10 mg/l






Concomitant palliative chemotherapy







Primary efficacy endpoints

The results are reported in Table 2. LBM assessed both by BIA (mean increase, +0.6 ± 2.4 kg) and DEXA (mean increase, +0.6 ± 2.7 kg) increased significantly (p < 0.0001). The proinflammatory cytokine TNF-α decreased significantly (mean decrease, −6.9 ± 11 pg/ml; p = 0.007). REE and fatigue did not change significantly, although, both of them showed a trend toward a decrease (p = 0.07 for REE and p = 0.09 for MFSI-SF score). Among patients with weight loss associated with high TNF-α levels (16 out of 18), ten showed an increase of LBM accompanied by a decrease of TNF-α levels, while two showed an increase of LBM without a concomitant decrease of TNF-α.
Table 2

Primary and secondary endpoints at baseline and after treatment



After treatment


Primary endpoints

LBM (kg)

BIA (n = 24)

45.4 ± 6.7

45.8 ± 6.6


DEXA (n = 18)

42.4 ± 6.6

43 ± 6.7


REE (Kcal/die)

1368 ± 405

1,324 ± 356



21.6 ± 25.2

15.8 ± 19.4


IL-6 (pg/ml)

22.9 ± 9.12

19.1 ± 12.8


TNF-α (pg/ml)

36.8 ± 14.1

29.9 ± 12.2


Secondary endpoints

Grip strength

20.8 ± 4.7

24 ± 5.5



5.7 ± 3.2

7 ± 2



432 ± 124

351 ± 169


GPx (IU/ml)

5,863 ± 3,185

7,674 ± 2,990



66.3 ± 19.9

75.3 ± 10.1



0.7 ± 0.3

0.7 ± 0.3



55.7 ± 19.9

60 ± 14.9



1.3 ± 0.77

0.8 ± 0.7



1.5 ± 0.67

1.2 ± 0.59


Objective clinical response



4 (16.67%)




20 (83.33%)


LBM lean body mass, BIA bioimpedance analysis, DEXA dual energy X-ray absorptiometry, REE resting energy expenditure, IL interleukin, TNF tumor necrosis factor, ROS reactive oxygen species, GPx glutathione peroxidase, EORTC QLQ-C30 European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-C30, EQ-5D Euro QL-5D, GPS Glasgow prognostic score, CR complete response, PR partial response, SD stable disease, PD progressive disease

aOne way analysis of variance (ANOVA). Results are considered significant for p < 0.05 (significant results are reported in bold)

Secondary efficacy endpoints

The results are reported in Table 2. Body weight and BMI did not change significantly. Grip strength increased significantly (p = 0.004), as well as the EORTC-QLQ-C30 and EQ-5D score (p = 0.024). GPS and ECOG PS decreased significantly (p = 0.0004 and p = 0.0023, respectively).

As for objective clinical response, SD was observed in four patients and PD in 20 patients at the end of treatment. Median progression-free survival (months) was 4 ± 1.2 (range 4–8). Median overall survival was 6 ± 2.7 (range 4–12).


There were no grade 3/4 adverse events. Grade 1/2 anemia was observed in three patients, neutropenia in two patients, and epigastralgia in one patient. The worst toxicity per patient is reported in Table 3. As for cardiac toxicity, neither symptomatic nor instrumental cardiac abnormalities were observed during treatment. Overall, patient compliance was very good, and no patient had to reduce the celecoxib dosage or interrupt treatment.
Table 3

Worst toxicity per patient


Grade 1/2

Grade 3/4










Gastrointestinal bleeding









Major cardiovascular eventsa



Instrumental pathological alterations (by electro- and/or echocardiography)



aMajor adverse cardiovascular events including CV death (including hemorrhagic death), nonfatal myocardial infarction, nonfatal stroke, hospitalization for unstable angina, revascularization, or hospitalization for a transient ischemic attack


The results of the present study show that COX-2 inhibitor celecoxib administered alone at 300 mg/day for 4 months induced an increase of LBM, a decrease of serum TNF-α, and a trend toward a decrease of fatigue symptom. Moreover, an improvement of grip strength and QL (EORTC-QLQ C-30) and a decrease of GPS and ECOG PS were observed. This strengthens the previously reported evidence that COX-2 inhibitors, such as celecoxib, are able to down-regulate TNF-α production/release and the dependent transcriptional activity and cellular pathways [35, 36], thus, influencing the inflammatory-driven signs and symptoms, such as GPS score.

The efficacy of celecoxib both on the inflammatory response mediators (cytokines and GPS) and on primary efficacy endpoint clinical/functional variable, LBM adds further evidence to the assumption that the core symptoms of cachexia, such as muscle wasting, fatigue, poor PS, and QL are systemic inflammation driven. Indeed, the improvement of a clinically significant functional symptom such as grip strength and improvement of QL should be considered a relevant finding directly dependent on the positive changes of LBM (Fig. 1).
Fig. 1

Main findings of the effects of celecoxib on cancer cachexia. REE resting energy expenditure, MFSI-SF multidimensional fatigue symptom inventory-short form, LBM lean body mass, BIA bioimpedentiometry, DEXA dual energy X-ray absorptiometry, TNFalpha tumor necrosis factor alpha, IL interleukin

Clinical trials on the effects of COX inhibitors used alone or in combination on cancer cachexia have been already published and should be compared with our trial. In a retrospective study on unselected weight-losing cancer patients, the use of the long-term COX inihibitor indomethacin was associated with an attenuation of resting energy expenditure and improved appetite paralleled by a decrease in CRP and erythrocyte sedimentation rate levels [37]. More recently, a randomized, placebo-controlled, pilot study carried out in a population of 11 cachectic patients with head and neck or gastrointestinal cancer showed a statistically significant increase in weight and BMI, as well as QL score in patients receiving celecoxib (200 mg twice daily) in comparison to the placebo arm [27]. Moreover, some combination regimens showed positive results on the primary features of CACS. In a randomized clinical trial on weight-losing patients with gastrointestinal cancer, ibuprofen combined with megestrol acetate appeared to reverse weight loss and to improve QL [38]. In a study by Cerchietti et al. [26], 15 patients with nonsmall cell lung cancer and evidence of CACS received a treatment with medroxyprogesterone acetate (500 mg twice daily) and celecoxib (200 mg twice daily) plus oral administration of food supplement for a period of 6 weeks. After treatment, 13 patients either had stable weight or gained weight. There were also significant differences in improvement from nausea, early satiety, fatigue, of appetite and PS. On the basis of these promising results, the same author later performed a randomized study on a cohort of 22 patients with advanced lung cancer who were assigned to receive either fish oil, 2 g tid plus placebo capsules bid (n = 12) or fish oil, 2 g tid plus celecoxib 200 mg bid (n = 10). All patients in both groups were orally administered with food supplements. After 6 weeks of treatment, patients receiving fish oil plus placebo or fish oil plus celecoxib showed significantly more appetite, less fatigue, and lower CRP values in comparison to baseline. Additionally, patients in the fish oil plus celecoxib group also improved their body weight and muscle strength compared to baseline values. Comparing both groups, patients receiving fish oil plus celecoxib showed significantly lower CRP levels and higher muscle strength and body weight than patients receiving fish oil plus placebo [39].

Among the combined approaches, one of the most intriguing is our open phase II trial [25, 30] whose aim was to test the safety and efficacy for clinical response, improvement of nutritional and functional variables, changes of laboratory variables, and improvement of QL, particularly fatigue, of an integrated treatment based on diet, oral pharmaconutritional support, and drugs in a population of advanced cancer patients with CACS. The combined regimen included celecoxib 200 mg/day plus polyphenols (400 mg), antioxidants (300 mg/day alpha-lipoic acid plus 2.7 g/day carbocysteine lysine salt plus 400 mg/day vitamin E plus 30,000 IU/day vitamin A plus 500 mg/day vitamin C), an EPA-enriched pharmaconutritional support, and 500 mg/day medroxyprogesterone acetate. The treatment duration was 4 months. Thirty-nine patients completed the treatment and were assessable. Body weight increased significantly from baseline as did LBM and appetite. There was a significant decrease of proinflammatory cytokines IL-6 and TNF-α, and interestingly, a negative correlation was found between LBM and IL-6 changes. As for QL, there was a marked improvement in the EORTC QLQ-C30, EQ-5DVAS, and fatigue. Therefore, the treatment proved to be effective and safe.

In comparison with the above-cited clinical trials, the present study is the first, to our knowledge, to show that an intervention with low dose (300 mg), midterm duration (4 months) of COX-2 inhibitor alone is able to induce a significant improvement of key symptoms of cancer cachexia among which LBM may be considered the most paradigmatic and clinically significant. Moreover, the treatment was safe without evidence of cardiac events with an optimal patient compliance.

The complete lack of toxicity of celecoxib in the present study is an encouraging finding considering the concerns which arose in 2004/2005 on the clinical use of COX-2 inhibitors because of increased cardiovascular risks leading to the withdrawal of rofecoxib and valdecoxib from the market. As regards celecoxib, the most recent FDA warnings state that “Celebrex may cause an increased risk of serious cardiovascular thrombotic events, myocardial infarction, and stroke. Patients with cardiovascular disease or risk factors for cardiovascular disease may be at greater risk.” The increased risk for the composite endpoint of cardiovascular death, myocardial infarction, or stroke includes only dosages ≥400 mg/day. Indeed, a large population-based study showed that the use of celecoxib at conventional doses for a period <180 days was not associated with any significant change in cardiovascular risk [40]. The present study confirms that celecoxib at a moderate dosage (200–300 mg/day) and for a limited duration (<6 months) does not increase cardiovascular or gastrointestinal risk. Based on the results of the present study, celecoxib as single agent showed an intriguing efficacy against CACS, and therefore, a phase III clinical trial maybe warranted to compare its efficacy and safety against other effective single agents and/or combination regimens.

Although the treatment of cancer cachexia, a multifactorial syndrome, is more likely to yield success with a multitargeted approach, in the present study, we were able to show that a treatment, such as celecoxib, addressing a single target, albeit very important as chronic inflammation, could have positive effects. Combining different treatments addressing the major multiple targets of CACS, we should be able to achieve an even more significant improvement of CACS symptoms.

Taking this into account, we have designed and are currently carrying out an ongoing phase III randomized clinical trial and have selected the single potentially most effective agents, although not including celecoxib, able to improve the “key” variables of CACS [41]. In conclusion, the results of the present study may encourage/warrant in the near future the use of celecoxib as a single agent in comparison with other single most effective agents against CACS.


The authors take full responsibility for the content of the paper but thank Ms. Anna Rita Succa for her assistance in editing the article.

Disclosure statement

The authors have no financial or personal relationships with other people or organizations that could inappropriately influence their work.

Copyright information

© Springer-Verlag 2009