Graft-versus-host disease Weight loss and malnutrition in ... - Nature

Bone Marrow Transplantation (2002) 29, 231–236  2002 Nature Publishing Group All rights reserved 0268–3369/02 $25.00 www.nature.com/bmt

Graft-versus-host disease Weight loss and malnutrition in patients with chronic graft-versushost disease DA Jacobsohn1, J Margolis2, J Doherty2, V Anders2 and GB Vogelsang2 1

Oncology and Pediatrics, The Johns Hopkins University School of Medicine, Baltimore, MD, USA; and 2Hematologic Malignancies, Department of Oncology, The Johns Hopkins University School of Medicine, Baltimore, MD, USA

Summary: It is well known that weight loss occurs in patients with chronic graft-versus-host disease. However, the severity and frequency of weight loss in this population have not been adequately described. Recent data suggest that a body-mass index (BMI) below 21.9 is an independent risk factor for mortality. In our analysis we have shown that out of 93 patients with cGVHD, 43% are malnourished as evidenced by a BMI less than 21.9 and 14% are severely malnourished (BMI less than 18.5). In addition, there is a clear trend showing that patients with active, ongoing cGVHD have lower BMIs (P = 0.02). Furthermore, we show that many symptoms thought to contribute to weight loss in patients with cGVHD, such as odynophagia and oral sensitivity, are not related to weight loss in our population. We conclude that, in all likelihood, unknown causes still exist that are responsible for weight loss in this group of patients. Elevated resting energy expenditure and elevated serum tumor necrosis factor-␣ are potential contributors to weight loss that will be tested in future studies. We also conclude that treating cGVHD aggressively may help reverse weight loss and malnutrition, which may be independent risk factors for mortality in this population. Bone Marrow Transplantation (2002) 29, 231–236. DOI: 10.1038/sj/bmt/1703352 Keywords: chronic graft-versus-host disease; weight loss; malnutrition; hematopoietic stem cell transplantation

Chronic graft-versus-host disease (cGVHD) is a major complication of allogeneic hematopoietic stem cell transplantation (SCT). CGVHD occurs in 27–50% of siblingmatched SCTs,1–3 42–72% of unrelated SCTs,4,5 and 46– 50% of G-CSF-stimulated allogeneic peripheral blood SCTs.6,7 It can lead to major disability, opportunistic infections, and even death in some patients.8,9 Weight loss has been recognized as a feature of Correspondence: Dr DA Jacobsohn, Johns Hopkins Hospital, Pediatric Oncology/CMSC 800, 600 North Wolfe Street, Baltimore, MD 21287, USA Received 3 July 2001; accepted 21 October 2001

cGVHD.10 Unfortunately, few studies in the literature have quantified the extent of weight loss in these patients. In 1990, Lenssen et al11 reported the incidence of weight loss and nutrition-related problems in 192 patients 1 year after allogeneic transplant. The data in this study showed that 33% of patients with extensive cGVHD and 19% of patients with limited cGVHD had weight loss.11 However, the degree of weight loss in each patient was not quantified and patients were only followed to 1 year. The conclusion of this study and the assumptions in most reviews of cGVHD is that weight loss in this group of patients is due to ongoing clinical symptoms such as dysphagia and oral sensitivity which directly decrease patients’ nutritional intake. Weight loss is a very significant problem and therefore understanding causes and mechanisms underlying weight loss in cGVHD could eventually lead to better treatment for this symptom. In general, wasting and cachexia are associated with severe physiologic, psychologic, and immunologic consequences, regardless of the cause.12–14 Cachexia has been associated with increased numbers of infections, decubitus ulcers, and even death.15–17 In patients with cancer cachexia, the degree of depletion of lean body mass and visceral protein is associated with reduced survival.18 This study was undertaken because we have seen a number of patients with cGVHD who are many years post SCT (5 or 10 years) and who are either losing weight or have been unable to regain weight since their transplant. While some of these patients have symptoms such as dysphagia that prevent nutritional intake, a significant number of patients with adequate nutritional intake have no apparent symptoms that can directly explain their weight loss. It has been hypothesized that patients with cGVHD have an altered metabolism leading to weight loss. In fact, one study recently reported that patients with extensive cGVHD have an abnormally elevated resting energy expenditure.19 Our aim in this study was two-fold. First, we wanted to quantify the degree of weight loss and malnutrition present in patients with cGVHD. Recently, low body-mass index (BMI) has become an important predictor of mortality in general medicine.20 We therefore felt that in addition to documenting weight change in our patients, determining BMI during ongoing cGVHD as an index of malnutrition

Weight loss in chronic GVHD DA Jacobsohn et al

232

would be a useful parameter to demonstrate the gravity of the problem. Second, we aimed at demonstrating the relationship between oral/gastrointestinal complaints such as dysphagia and oral sensitivity, and degree of weight loss. Our hypothesis is that while these symptoms may explain weight loss in some patients, there are still unknown factors contributing to weight loss in cGVHD.

used: (1) no symptoms, (2) food sensitivity, (3) pain, requiring narcotics, and (4) unable to eat. Malabsorption was defined as laboratory-proven steatorrhea. Nutritional supplementation was defined as any additional nutrition supplement to a standard diet being used by the patient at the time of the clinic visit. While the vast majority of nutritional supplementation consisted of formulas such as Vivonex and Ensure, a few patients received feeds via gastric tubes.

Methods Patients In 1994 a multidisciplinary GVHD clinic was established at the Johns Hopkins Oncology Center. This clinic’s treatment recommendations focus not only on specific therapy for GVHD but also on rehabilitation, nutritional, and psychosocial problems. This study is a retrospective chart review of 104 patients referred to this GVHD clinic between 1994 and 1998 in whom a diagnosis of cGVHD was confirmed. The main objective of the study was to analyze malnutrition and weight loss patterns in this group of patients as well as to comprehend what factors may be contributing to these patterns. For patients’ charts that were missing data, we contacted the patient or the patient’s oncologist to obtain this information.

Body-mass index BMI was calculated as the weight in kilograms (at clinic visit) divided by the square of the height in meters. Patients were stratified according to the classification used in the study by Calle et al20 in which more than 1 million adults in the United States were followed prospectively to assess the relationship between mortality and BMI. In this study, a higher risk of death was found to occur in those with a BMI below 21.9, with an even higher risk in those with a BMI below 18.5 (RR of death 1.26 for men and 1.36 for women). In addition, those with a BMI above 25.0 had progressively higher risk of mortality as BMI increased. Therefore, we divided our patients into four categories that best simulated the mortality risk data presented by Calle.

Study definitions Diagnosis of cGVHD by our clinic was made on the basis of both clinical and histologic criteria for the skin and other affected sites, as previously described.21–23 The histologic diagnosis was made using established criteria.23,24 Patients brought slides of a biopsy of the affected organ, or a biopsy was performed at our institution; all samples were examined by pathologists at the Johns Hopkins Hospital. Time of therapy for cGVHD was defined as the number of years that had elapsed between the time the patient was diagnosed with cGVHD and started on therapy by the referring physician, and the date that the patient was first seen at our clinic. Patients were identified as having active cGVHD if they had a biopsy confirming cGVHD and had evidence of progressive or persistent disease over the past 3–6 months (based on the patient’s and referring physician’s reports). Patients were identified as having inactive cGVHD if they had a biopsy showing cGVHD but their disease had been completely stable over the past 3–6 months. A period of 3–6 months was chosen because we had previously observed at our institution that the majority of patients who have had no improvement (as well as no progression) do not improve if further maintained on immunosuppressive therapy.25 Skin cGVHD was defined as the presence of biopsyproven lichenoid or sclerodermatous cGVHD. For the categories of odynophagia, dysphagia, nausea, anorexia and abdominal pain – patients were specifically asked, in lay language, if they had experienced these symptoms and the answer was recorded as yes/no. For the oral sensitivity symptom, a 4-grade category was Bone Marrow Transplantation

Weight change Percent weight change was calculated as the difference between weight at the clinic visit and weight pre-SCT, divided by weight pre-SCT. Based on previous literature, patients were stratified into three different groups: (a) weight gain or ⬍5% weight loss, (b) 5–10% weight loss, and (c) ⬎10% weight loss. In a study of 84 patients with metastatic renal cancer, weight loss of greater than 10% was predictive of decreased survival.26 Wallace et al27 found that in 247 males over the age of 65, an involuntary weight loss of greater than 4% was an independent predictor of increased mortality.

Statistical analysis

␹2 tests were used to compare the frequency distributions of symptoms/diagnoses in each weight group category. The trend test was used for comparing proportions with three or more levels (eg in the analysis using stratified BMI). Associations between diagnoses/symptoms were assessed by odds ratios (OR) estimated by a univariate multinomial logistic regression model. We repeated the analyses in a multivariate model to assess whether there was any confounding. Since the associations were almost identical in both models and most of the odds ratios were not ‘statistically significant’, we present the unadjusted results. The ttest was used to assess the difference between two means. All statistical analyses were carried out with Stata 6.0 software (College Station, TX, USA).


Results

Table 2 their visit

Patient characteristics One-hundred and four adult patients referred to the Johns Hopkins Oncology Center for management of cGVHD from January 1994 to July 1998 had the diagnosis of cGVHD confirmed. Out of these, 93 of them had enough information (height as well as weight) to calculate BMI. The characteristics of these patients are listed in detail in Table 1. The median age of patients was 37 years, with a range of 17 to 64 years. Individuals under the age of 17 years were excluded since these patients should be gaining weight as part of normal growth and this issue would confound our analysis. There was an almost equal balance between males and females. The median time since SCT was 2.4 years and the median time on therapy for cGVHD was 1.7 years. The most common type of SCT was an allogeneic, matched-sibling donor transplant and the most common indication was chronic myelogenous leukemia. The sites of cGVHD for these patients are detailed in Table 2.

Table 1

Patient characteristics

Number of patients

93

Age at visit, years Median Range

37 17–64

Gender

45 F/48 M

Time since stem cell transplant, years Median Range

2.4 0.33–12.9

Time on therapy for cGVHD,a years Median Range Previous immunosuppressive cGVHD therapy Prednisone/methylprednisolone Cyclosporine/FK506 Azathioprine Mycophenolate mofetil Type of stem cell transplant Matched-sibling donor Matched-unrelated donor 1-antigen mismatched donor 2-antigen mismatched donor Matched-sibling PBSC Matched-sibling, followed by DLI Autologous, followed by matched-sibling Indications for bone marrow transplant Chronic myelogenous leukemia Acute myelogenous leukemia Aplastic anemia Acute lymphoblastic leukemia Myelodysplastic syndrome Multiple myeloma Non-Hodgkin’s lymphoma Hodgkin’s disease Chronic lymphocytic leukemia

1.7 0–12.6 86 81 61 6

(92%) (87%) (65%) (6%) 59 20 6 1 2 4 1 42 23 6 6 4 4 4 2 2

a Number of patients who had received medication for at least 6 months. cGVHD = chronic graft-versus-host disease; PBSC = peripheral blood stem cell transplant; DLI = donor lymphocyte infusion.

Site of chronic graft-versus-host disease in 93 patients at

Site

No.

%

Skin Sclerodermatous Lichenoid Oral Liver Musculoskeletal

52 10 32 6 17

56 11 34 6 15

233

Body-mass index BMI at the time of cGVHD referral was calculated for each patient. Patients were also categorized as having active or inactive cGVHD. The detailed results of this analysis are presented in Table 3. The main results to highlight are that 43% of our patients were in the malnourished category: 14% were in the severe category (BMI ⬍18.5) and 29% were in the moderate category (BMI 18.5–21.9). In addition, there is a clear trend indicating that patients with active cGVHD have significantly lower BMIs when compared to patients with inactive cGVHD (P for trend difference = 0.02). In fact, out of 13 patients who presented with a BMI less than 18.5, 12 (92%) were in the active cGVHD category and only one (8%) was in the inactive cGVHD category. Weight change analysis In 71 of the 93 patients listed above, we were able to calculate weight change since SCT. The percent weight change was calculated for each patient as described in the Methods section. Two patients were deleted from the analysis since their weight change significantly outlied the rest of the data (they were beyond the interquartile ranges on a boxplot for weight change) and thus the concern existed that these numbers were incorrect. Therefore, a total of 69 patients was used in this analysis. To determine if active cGVHD had a significant influence on weight loss, we separated our 69 patients into active/inactive cGVHD status and looked for a difference in mean weight loss in each group. The details of these

Table 3 Body-mass index of 93 patients according to chronic graftversus-host disease classificationa BMI stratum

Inactive cGVHD n (%)

Active cGVHD n (%)

⬍18.5 18.5–21.9 22.0–24.9 ⬎24.9

1 7 3 12

12 20 24 14

Total

23 (100)

(4) (30) (13) (53)

(17) (29) (34) (20)

70 (100)

Total n (%)

13 27 27 26

(14) (29) (29) (28)

93 (100)

P for trend difference = 0.02.

a

Bone Marrow Transplantation


234

loss by 3.22 (0.44–23.6) times. Finally, age, time since SCT, and duration of therapy for cGVHD had no effect on weight loss.

Table 4 Weight loss pattern in 69 patients with chronic graft-versushost disease according to active or inactive disease

Number of patients Mean % weight change (95% CI) Median % weight change Range % weight change a

Active

Inactive

50 −8.7% (−12.4% → −5.0%) −9.3% −33.7% → +22.1%

19 −1.4% (−7.3% → +4.5%)a +0.7% −29.5% → +16.4%

Discussion In our study, we have examined the incidence of malnutrition and weight loss in patients with cGVHD, as well as potential factors that may contribute to these. Calculating BMI for each patient allowed us to stratify patients into malnutrition severity categories. Almost half of the patients in our study were malnourished and 14% of the patients were severely malnourished. In addition, the majority of the severely malnourished patients had active, ongoing cGVHD. These data are alarming in the light of Calle’s study showing an increased risk of mortality in patients with a BMI below 21.9.20 Our data also suggest that successful therapy of cGVHD helps decrease malnutrition since patients with inactive cGVHD had lower rates of malnutrition as compared to patients with active cGVHD. Thus, attempting to reverse malnutrition with medical and nutritional management should be a priority in a cGVHD consultation since malnutrition may in itself be an independent risk factor for death. A further study will focus on defining the potential association between malnutrition and mortality in this specific group of patients. We also examined the relationship between weight loss in cGVHD patients and gastrointestinal complaints that could be contributing to weight loss. Interestingly, symptoms such as nausea, anorexia, and oral sensitivity do not appear to be related to weight loss. The presence of dysphagia does appear to be related to weight loss in cGVHD patients. This finding is not surprising since there are specific esophageal abnormalities, such as webs, strictures, and lack of peristalsis, that have been reported in patients with cGVHD.28 The presence of abdominal pain showed some positive relationship with weight loss, although this finding was not statistically significant. Perhaps these patients have ‘gastrointestinal cGVHD’, although this is not a finding that

P = 0.04 for the difference of means.

results are shown in Table 4. Active cGVHD patients had a mean weight loss of 8.7% while inactive cGVHD patients had a mean weight loss of 1.4% (P = 0.04). The following factors were examined as potentially being related to weight change: active vs inactive cGVHD, dysphagia, odynophagia, nausea, anorexia, abdominal pain, oral sensitivity, malabsorption, skin cGVHD, and nutritional supplementation. Table 5 displays the incidence of each symptom/diagnosis in each weight change category. The only statistically significant factor is nutritional supplementation, with the majority of patients in the ⬎10% weight loss category being on supplementation. Otherwise, dysphagia approached statistical significance, with a higher percentage of patients in the ⬎10% and 5–10% weight loss categories having dysphagia (45% and 63%, respectively) as compared to those in the baseline category (25%). Furthermore, univariate multinomial logistic regressions were performed to calculate the odds ratios of each symptom/diagnosis on the incidence of weight loss. The results of this analysis are detailed in Table 6. Certain results in this analysis deserve highlighting. Active cGVHD increased the risk of ⬎10% weight loss by 2.9 (0.86–9.48) times and of 5–10% weight loss by 1.8 (0.31–10.38) times. Dysphagia increased the risk of ⬎10% weight loss by 2.43 (0.83–7.24) times and of 5–10% weight loss by 4.95 times (0.97–25.8). Abdominal pain increased the risk of ⬎10% weight loss by 2.5 (0.57–11.1) times and of 5–10% weight Table 5

Symptoms in 69 patients with chronic graft-versus-host disease stratified by weight category Weight gain/⬍5% loss n (%)a

5–10% weight loss n (%)a

⬎10% weight loss n (%)a

Total n (%)a

Active cGVHD Dysphagia

20 (63) 8 (25)

6 (75) 5 (63)

24 (83) 13 (45)

Odynophagia Nausea Anorexia Abdominal pain Oral sensitivityb Malabsorption Skin cGVHD Nutritional supplementation

5 14 8 3 20 0 25 9

1 4 3 2 4 0 5 3

4 11 7 6 21 1 20 21

50 (72) 26 (38) (P = 0.08) 10 (14) 29 (42) 18 (26) 11 (16) 45 (65) 1 (1) 50 (72) 33 (48) (P ⬍ 0.001)

Total a

(16) (44) (25) (9) (63) (0) (78) (28) 32

(13) (50) (38) (25) (50) (0) (83) (50) 8

(14) (38) (24) (21) (72) (3) (69) (72) 29

Percentage of patients reporting/having that symptom out of total number of patients in that weight category. Dichotomized in this analysis as yes/no. All P values testing differences among groups by chi-square are ⬎0.1 unless specified.

b


69 (100%)


Table 6 Univariate logistic regressions examining potential factors related to weight loss in 69 patients with chronic graft-versus-host disease

Active vs inactive cGVHD Odynophagia Dysphagia Nausea Anorexia Oral sensitivitya Abdominal pain Skin cGVHD Ageb Time since BMTb Time of therapy for CGVHDb

Odds ratio 5–10% weight loss

Odds ratio ⬎10% weight loss

1.80 (0.31–10.38)

2.88 (0.86–9.48)

0.77 4.95 1.28 1.80 0.95 3.22 0.47 1.02 1.02 1.08

0.86 2.43 0.79 0.95 1.23 2.53 0.63 0.98 1.21 1.20

(0.08–7.69) (0.97–25.8) (0.27–6.05) (0.35–9.28) (0.47–1.93) (0.44–23.6) (0.09–2.46) (0.95–1.10) (0.70–1.48) (0.76–1.55)

(0.21–3.60) (0.83–7.24) (0.28–2.18) (0.30–3.07) (0.79–1.93) (0.57–11.1) (0.20–1.97) (0.93–1.03) (0.96–1.52) (0.96–1.51)

a Expressed as odds ratio for each increasing grade in severity of oral sensitivity. b Expressed as odds ratio for each additional year.

has been adequately defined or documented. Our group is in the process of doing a comprehensive review of cGVHD gastrointestinal complications to understand better the clinical findings and histology that may represent ‘gastrointestinal cGVHD’. An interesting finding is that the majority of patients with severe weight loss were on nutritional supplementation at the time of the cGVHD consultation. While this may seem like an obvious relationship, it could potentially signify that patients with severe weight loss are not gaining weight after having lost it despite being on nutritional supplementation. This is only a hypothesis at this point, since we do not have adequate information such as length and type of supplementation, or weights at multiple timepoints, to draw any conclusions. We hypothesize that there are still unknown factors leading to weight loss in cGVHD. It is likely that patients with extensive cGVHD are in a state of tumor necrosis factor␣ (TNF-␣)-dependent catabolism. Patients that develop cGVHD have elevated levels of cytokines, in particular, TNF-␣.29 TNF-␣ is involved in cachexia induced by chronic illnesses.30,31 Furthermore, certain cancer patients are in a state of catabolism as documented by elevated resting energy expenditures (REE).32 In fact, one small study has demonstrated REE in patients with extensive cGVHD.19 The main limitation with this study is that the controls used were drawn from the general population, and thus, theoretically, one cannot conclude that the elevated REE is secondary to cGVHD as opposed to another postSCT complication. In an effort to further explore the above hypothesis, we plan to launch a clinical trial in which we will measure REE, protein oxidation rates, and cytokines in cGVHD patients with weight loss and post-SCT controls. We will also administer infliximab, an anti-TNF monoclonal antibody, to cGVHD patients with weight loss in an attempt to reverse their weight loss and altered metabolism. There are certain limitations to our study that deserve mention. First of all, much of the study is retrospective in

nature. Therefore, it is sometimes difficult to assess severity and length of symptoms. Our case encounter forms, however, prompt the physician to ask about nutrition-related problems and therefore we feel the data are quite complete. Furthermore, patients and their physicians were called if there were pieces of information missing. Another potential limitation is that the weight change pattern is difficult to assess since we are evaluating two weights for the patients – current visit weight and pre-SCT weight. One might argue that we missed a patient at their weight peak, for example. While this is a true concern, we feel that the weight change data are valid when taken as a whole. We analyzed each symptom by current BMI and the results were almost identical to those when using percent weight change (data not presented). Therefore, a large percent weight loss correlates very well with low BMI, which suggests that patients who have lost a lot of weight are malnourished regardless of their starting weight pre-SCT or a potential missed peak in weight. Finally, it is worth mentioning that while weight loss and malnutrition are significant problems in our population, almost half of the patients had weight gain or less than 5% weight loss, and more than half of these had active, ongoing cGVHD. It is conceivable that steroid therapy in these patients led to weight gain from fluid retention and adipose tissue accumulation. Our future study will allow us to calculate lean body mass for each patient and distinguish between adequate and inadequate forms of weight gain. In conclusion, we have shown that weight loss and malnutrition are very significant problems in patients with cGVHD. Furthermore, it appears that those with active, ongoing disease have a higher incidence of weight loss and malnutrition. Therefore, treating cGVHD aggressively may help reverse weight loss and malnutrition, which in general are independent risk factors for mortality. However, whether weight loss and malnutrition increase mortality in cGVHD patients remains unknown. This association will be studied in detail once we have accrued more follow-up time with this cohort of patients. Given our findings, it is likely that unknown causes still exist that are responsible for weight loss in this group of patients. Future studies will be aimed at understanding these causes and devising more specific ways to treat weight loss and malnutrition in this population.

235

References 1 Lucarelli G, Galimberti M, Polchi P et al. Bone marrow transplantation in patients with thalassemia. New Engl J Med 1990; 322: 417–421. 2 Gaziev D, Polchi P, Galimberti M et al. Graft-versus-host disease after bone marrow transplantation for thalassemia: an analysis of incidence and risk factors. Transplantation 1997; 63: 854–860. 3 Lazarus HM, Rowe JM. New and experimental therapies for treating graft-versus-host disease. Blood Rev 1995; 9: 117– 133. 4 Beatty PG, Hansen JA, Longton GM et al. Marrow transplantation from HLA-matched unrelated donors for treatment of hematologic malignancies. Transplantation 1991; 51: 443– 447. Bone Marrow Transplantation


236

5 Morton AJ, Anasetti C, Gooley T et al. Chronic graft-versushost disease (GVHD) following unrelated donor transplantation. Blood 1997; 90: 590a (Abstr. 2623). 6 Bensinger WI, Martin PJ, Storer B et al. Transplantation of bone marrow as compared with peripheral-blood cells from HLA-identical relatives in patients with hematologic cancers. New Engl J Med 2001; 344: 175–181. 7 Blaise D, Kuentz M, Fortanier C et al. Randomized trial of bone marrow versus lenograstim-primed blood cell allogeneic transplantation in patients with early-stage leukemia: a report from the Societe Francaise de Greffe de Moelle. J Clin Oncol 2000; 18: 537–546. 8 Atkinson K, Farewell V, Storb R et al. Analysis of late infections after human bone marrow transplantation: role of genotypic nonidentity between marrow donor and recipient and of nonspecific suppressor cells in patients with chronic graft-versus-host disease. Blood 1982; 60: 714–720. 9 Atkinson K. Chronic graft-versus-host disease. Bone Marrow Transplant 1990; 5: 69–82. 10 Flowers ME, Kansu E, Sullivan KM. Pathophysiology and treatment of graft-versus-host disease. Hematol Oncol Clin North Am 1999; 13: 1091–1112. 11 Lenssen P, Sherry ME, Cheney CL et al. Prevalence of nutritionrelated problems among long-term survivors of allogeneic marrow transplantation. J Am Diet Assoc 1990; 90: 835–842. 12 Chandra RK. Nutrition, immunity, and infection: present knowledge and future directions. Lancet 1983; 1: 688–691. 13 Haydock DA, Hill GL. Improved wound healing response in surgical patients receiving intravenous nutrition. Br J Surg 1987; 74: 320–323. 14 Lennmarken C, Larsson J. Skeletal muscle function and energy metabolites in malnourished surgical patients. Acta Chir Scand 1986; 152: 169–173. 15 Pinchcofsky-Devin GD, Kaminski MV Jr. Correlation of pressure sores and nutritional status. J Am Geriatr Soc 1986; 34: 435–440. 16 Silver AJ, Morley JE, Strome LS et al. Nutritional status in an academic nursing home. J Am Geriatr Soc 1988; 36: 487–491. 17 Rudman D, Feller AG, Nagraj HS et al. Relation of serum albumin concentration to death rate in nursing home men. J Parenter Enteral Nutr 1987; 11: 360–363. 18 Nixon DW, Heymsfield SB, Cohen AE et al. Protein-calorie undernutrition in hospitalized cancer patients. Am J Med 1980; 68: 683–690.


19 Zauner C, Rabitsch W, Schneeweiss B et al. Energy and substrate metabolism in patients with chronic extensive graft-versus-host disease. Transplantation 2001; 71: 524–528. 20 Calle EE, Thun MJ, Petrelli JM et al. Body-mass index and mortality in a prospective cohort of US adults. New Engl J Med 1999; 341: 1097–1105. 21 Shulman HM, Sullivan KM, Weiden PL et al. Chronic graft-versus-host syndrome in man. A long-term clinicopathologic study of 20 Seattle patients. Am J Med 1980; 69: 204–217. 22 Tutschka PJ, Beschorner WE, Hess AD, Santos GW. Cyclosporin-A to prevent graft-versus-host disease: a pilot study in 22 patients receiving allogeneic marrow transplants. Blood 1983; 61: 318–325. 23 Farmer ER. The histopathology of graft-versus-host disease. Adv Dermatol 1986; 1: 173–188. 24 Saurat JH, Gluckman E, Bussel A et al. The lichen planuslike eruption after bone marrow transplantation. Br J Dermatol 1975; 93: 675–681. 25 Wingard JR, Piantadosi S, Vogelsang GB et al. Predictors of death from chronic graft-versus-host disease after bone marrow transplantation. Blood 1989; 74: 1428–1435. 26 Mani S, Todd MB, Katz K, Poo WJ. Prognostic factors for survival in patients with metastatic renal cancer treated with biological response modifiers. J Urol 1995; 154: 35–40. 27 Wallace JI, Schwartz RS, LaCroix AZ et al. Involuntary weight loss in older outpatients: incidence and clinical significance. J Am Geriatr Soc 1995; 43: 329–337. 28 McDonald GB, Sullivan KM, Schuffler MD et al. Esophageal abnormalities in chronic graft-versus-host disease in humans. Gastroenterology 1981; 80: 914–921. 29 Remberger M, Ringden O, Markling L. TNF alpha levels are increased during bone marrow transplantation conditioning in patients who develop acute GVHD. Bone Marrow Transplant 1995; 15: 99–104. 30 Oliff A, Defeo-Jones D, Boyer M et al. Tumors secreting human TNF/cachectin induce cachexia in mice. Cell 1987; 50: 555–563. 31 Tracey KJ, Wei H, Manogue KR et al. Cachectin/tumor necrosis factor induces cachexia, anemia, and inflammation. J Exp Med 1988; 167: 1211–1227. 32 Fredrix EW, Wouters EF, Soeters PB et al. Resting energy expenditure in patients with non-small cell lung cancer. Cancer 1991; 68: 1616–1621.