C-reactive Protein Levels and Decrease of Albumin Levels in Hospitalized Elderly Patients with Community-Acquired Pneumonia

Transcription

1 ORIGINAL ARTICLE C-reactive Protein Levels and Decrease of Albumin Levels in Hospitalized Elderly Patients with Community-Acquired Pneumonia Kuntjoro Harimurti, Siti Setiati ABSTRACT Aim: to obtain: (1) the correlation between initial CRP level on admission with the decrease of albumin level during hospitalization, (2) the mean difference in initial CRP level between the groups of patients with and without decrease of albumin level during hospitalization, and (3) the risk difference of decreasing albumin level in patients with high CRP levels on the admission compared to whom with low CRP level on the admission, in hospitalized elderly patients with CAP. Methods: a prospective cohort study were conducted on 23 hospitalized elderly patients with CAP. Subjects with diseases and conditions that could interfere with CRP and albumin level besides pneumonia infection were excluded. The patient s CRP level was measured upon the initiation of the study, while the patient s albumin level was measured on the rst and fth day of hospitalization to observe changes that took place during 5 days of hospitalization. Pearson s correlation test, independent t-test, and chi-square test were used to answer the objectives of the study. Results: we found that there were negative correlation between the initial CRP level and the percentage of albumin level decrease during 5 days of hospitalization (r=-0.442, p= 0.035) and signi cance difference in the mean initial CRP level between patients with and without decreasing albumin level (mean difference mg/l, 95%CI to mg/l; P 0.026). The risk difference of decreasing albumin level during hospitalization between patients with high and low initial CRP levels did not attaint statistical signi cance (RR %CI 0.26 to 29.07; p=0.621). Conclusion: in hospitalized elderly patients with community-acquired pneumonia with high initial CRP levels tend to experience a decrease in albumin level during hospitalization. Key words: CRP, albumin, community-acquired pneumonia, elderly. Division of Geriatric, Department of Internal Medicine, Faculty of Medicine of the University of Indonesia/Cipto Mangunkusumo National Central General Hospital, Jakarta INTRODUCTION Community-acquired pneumonia (CAP) is a common health problem in elderly patients. Some study reported as well as epidemiological and clinical data demonstrate that the incidence, prevalence, and the mortality rate of CAP increase as a person gets older. 1-4 One of the factors that facilitates the development of pneumonia in an elderly person and acts as a negative prognosis factor and predictor of death in persons with pneumonia, is low albumin level. 5-8 The role of hypoalbuminemia in causing a bad course of disease is by far more related to malnutrition, which is often found in an elderly. 9 However, decreased albumin levels during acute infection is also directly caused by the underlying in ammation process, to some extent, could provide an illustration of the severity of infection/in ammation Because of this characteristic during acute in ammation, albumin often mentioned as a negative acute-phase protein. 16,17 At the same time the level of C-reactive protein (CRP) elevates, as a positive acute-phase protein that have been recognized and widely studied in various types of acute infection, includes pneumonia The relationship between albumin and CRP level in in ammation has been widely investigated, but mostly only in various chronic diseases, i.e. in patients with endstage renal disease (ESRD) undergoing hemodialysis and the elderly population assumed to have various chronic in ammation. 24 In patients with acute infection, especially pneumonia, there has only been one study that observed the correlation between albumin with CRP and other acute in ammatory markers at the beginning of hospitalization in hospitalized CAP patients. 8 There have not been any studies that particularly investigate the relationship between initial CRP level as a representation of the severity of underlying acute infection with the decrease in albumin level during hospitalization. Based on the result of previous studies and the 13

2 Kuntjoro Harimurti, et al pathophysiology of changes in albumin and CRP level in acute infection, we assume that high initial CRP levels in elderly patients hospitalized with CAP is related to decreased albumin level during hospitalization. Therefore, we designed this study, the objectives of which are to obtain: (1) the correlation between initial CRP level with the extent of decrease in albumin level during hospitalization, (2) the mean difference in initial CRP level between patients with and without decreased albumin levels during hospitalization, and (3) the increase in risk of decreasing albumin level during hospitalization in patients with high initial CRP levels compared to patients with low initial CRP levels, in hospitalized elderly patients with CAP. The results of the study are expected to increase awareness about the incidence of decrease albumin level, along with its consequences, in patients with high CRP levels at the beginning of their hospitalization. METHODS A prospective cohort study was conducted in elderly patients (aged 60 years or more), diagnosed with CAP and hospitalized at Cipto Mangunkusumo Hospital, Jakarta. The initial CRP level and changes in albumin level after 5 days of hospitalization were measured. Patients with diseases or conditions other than CAP that could interfere with the albumin and CRP level was excluded from the study, i.e. patients with congestive heart failure functional class III or IV according to NYHA classi cation, chronic renal diseases with < 30 ml/minute of creatinine clearance, severe dehydration, ascites and/or anasarca oedema, liver cirrhosis, proteinuria (of 2+ or more, or 300 mg or more in 24 hours), stadium III or IV pressure ulcers, autoimmune diseases with or without immunosuppresant therapy, terminal stage malignancy, acute myocardial infarction (STEMI and Non-STEMI), moderate or severe immobilization for more than 3 days, and the presence of infections other than communityacquired pneumonia. CRP level was only measured upon admission, using Dade-Behring high-sensitivity CRP immunonephelometry method. Albumin level was measured twice, at the beginning of hospitalization and on the fth day of stay, using bromcresol green method. Initial CRP level was correlated with changes in albumin level (in percent) during the ve days of hospitalization, and the mean difference of initial CRP level was calculated between patients with and without decrease in albumin level during the 5 days of hospitalization (a patient is considered to have a decrease in albumin level if the albumin level decreases >10% from the initial albumin level). The relative risk of decrease albumin level 14 Acta Med Indones-Indones J Intern Med during the 5 days of hospitalization was calculated in 2 groups of patient based on the initial CRP level (high CRP level, > 20 mg/l and low CRP level, < 20 mg/l). During hospitalization, the patients received antibiotics therapy and diet (including protein diet) as indicated, while patients who received albumin infusion during hospitalization were excluded from the study. Analysis were conducted on the data obtained as follows: (1) Pearson correlation test on initial CRP level and percentage of changes in albumin level during hospitalization as numeric variables; (2) independent t-test on the mean difference of initial CRP level between patients with and without decrease in albumin level; and (3) chi-square test to obtain the relative risk of albumin level decrease during hospitalization in patients with high initial CRP levels compared to patients with low initial CRP levels. All descriptions and analysis were conducted using the SPSS version 10.0 for Windows computer program. The study protocol had obtained ethical clearance from the Ethical Research Committee in Faculty of Medicine, University of Indonesia, Jakarta. RESULTS From March to June 2005, at Dr. Cipto Mangunkusumo Hospital, we found 26 elderly patients diagnosed with community-acquired pneumonia who fulfilled the subject selection criteria for the study, consisting of 13 male and 13 female subjects. Out of the 26 patients, only 23 patients completed the study until the fth day of hospitalization (two patients were discharged and one patient was died before ve days of observation). The three patients that did not complete the study were not calculated in cohort analysis, but the data obtained at the initiation of study are still included in the study subject characteristics. Basic laboratory test results, and the level of initial CRP and albumin level are presented in Table 1. After five days of hospitalization, seven out of 23 patients demonstrated a 10% or more decrease in albumin level, and an overall significant decrease was found in the mean albumin level on the fth day compared to the rst day measurement (3.24 g/dl to 3.05 g/dl, P 0.004). Table 2 shows the comparison of some clinical and laboratory variables (other than CRP level) at the beginning of hospitalization between patients with and without reduced albumin level during hospitalization. The data of the rst-day CRP level from 23 patients who completed the study showed abnormal distribution. After data transformation for correlation analysis, the rst-day CRP level showed negative correlation with

3 Vol 39 Number 1 January - March 2007 C-reactive Protein Levels and Decrease of Albumin Levels Figure 1. The correlation between the square root of CRP level on day-1 and reduction in albumin level after 5 days of hospitalization decreasing albumin level in patients with high initial CRP level compared to patients with low initial CRP level was 2.12 (95%CI 0.26 to 29.07; P 0.621). DISCUSSION percentage of decrease in albumin level during 5 days of hospitalization (r= P 0.035; Figure 1). We have mentioned above that a 10% or more decrease of albumin level was found in 7 out of 23 subjects who completed the study. When those seven patients were gathered as one group (group with albumin level decrease) and the rest were considered as one other group (group without albumin level decrease) using t-test in the independent data group, after normalization of data distribution, we found a signi cant difference in the mean rst-day CRP levels between the two groups (mean difference mg/l, 95% CI to mg/l; P 0.026). The relative risk (RR) of This is the rst study to investigate the relationship between CRP level with reduction in albumin level during acute infection in hospitalized elderly patients. Community-acquired pneumonia was chosen as a model of acute infection being studied, for two reasons: (1) the relatively high incidence, prevalence, and mortality of pneumonia in elderly population require attention to the factors that might contribute to the occurrence and course of the disease in the elderly,5,6 and (2) the severity of infection and in ammation quanti ed with CRP level, for the purpose of studying the effect to albumin level decrease in this study, have often been proven to be representative for pneumonia in all age groups, including the elderly The mean initial CRP level in elderly patients with pneumonia in this study was greatly increased (Table 1), and this is in line with results from other studies.19,20 This increase, although also found in other bacterial infections, is a sensitive clinical marker and could be used to make diagnosis and it re ects the severity of pneumonia infection, including in the elderly.18,19 The increase in CRP level in acute infection is stimulated by interleukin-6 (IL-6) as a proin ammation mediator, and also contributes to complement activation and interacts with the humoral and cellular immune system.11,16,17 Experimental studies also show that CRP has protective effects on lung infection (especially the ones caused by Streptococcus pneumoniae), because it is also expressed in the alveoli.25,26 15

4 Kuntjoro Harimurti, et al On the other hand, a signi cant decrease in mean albumin level was found on the fth day of hospitalization. Hedlund et al. found that serum albumin is decreased until the sixth day of hospitalization, and then it increases again on the ninth day, in patients with CAP. 8 Sullivan et al. in their prospective study found lower albumin level upon patient discharge from the hospital (mean 29.1 g/l) compared to the admission (mean 36.6 g/l), in hospitalized elderly patients. 9 Reduction in albumin level during hospitalization in patients hospitalized due to acute conditions seems to be greatly in uenced by the severity of underlying infection/ in ammation, although the role of nutrition could not be eliminated. This role of infection/in ammation that we investigated in this study. From the correlation study, a negative correlation was found between high CRP level with percentage of albumin level decrease during 5 days of hospitalization, although the correlation was not too strong. (Figure 1) In other words, the higher the patient s initial CRP level, the greater the decrease in albumin level during ve days of hospitalization. Two studies that correlates CRP and albumin level at the same time showed almost similar results. The study by Hedlund et al. found weak but signi cant negative correlation (r = P ) between CRP and albumin levels at the admission in patients with hospitalized community-acquired pneumonia. 8 A study in patients with end-stage renal disease having hemodialisis also found similar results (r = P <0.001). 21 The correlation between CRP and albumin level is interesting to observe, because both are acute-phase proteins whose serum concentration changes during acute infection/ in ammation. Since CRP is a positive acute-phase protein, while albumin is a negative acutephase protein, the correlation between both is expected to be negative. This study and two other studies mentioned above showed a negative correlation, although the correlation tends to be weak. However, by considering the rate of synthesis, half life, and greatly different responses of each acute-phase proteins in an infection/ in ammation process, 16 it seems to be dif cult to obtain the correlation between CRP and albumin level in parallel time. Besides, previous nutritional status in uences initial albumin level to varying degrees. 8,9,27,28 Therefore, the judgment to correlate initial CRP level with albumin level reduction during the rst few days of hospitalization that done in this study is more acceptable, instead of just correlating CRP and albumin level at the same time. The second nding of this study is the signi cant mean difference in initial CRP level between the group of Acta Med Indones-Indones J Intern Med patients with and without albumin level decrease during hospitalization. The rst group shows signi cantly higher initial CRP level compared to the second group (mean difference mg/l). This result is supports the result of the experimental studies that showed that beside increased of CRP level, increasing concentrations of proin ammation cytokines (especially IL-6) in the in ammation process inhibits albumin synthesis in the hepatocyte, 11,12 as well as increases albumin catabolism and redistribution to the extravascular compartment, 12,15,29 with the end result is reduction in circulatory albumin level. Since both are affected by increased IL-6 during acute in ammation, a decrease in albumin level occurs simultaneously with an increase in CRP level, but at a different rate. 11,16 An increase in CRP level precedes a decrease in albumin level, so a high initial CRP level as a marker for severity of infection could serve as predictor of albumin level decreased during hospitalization (especially on the rst few days of hospitalization). In adddition, from data showed in Table 2, the other infection markers (leukocyte count and ESR) are higher in patients with decreased albumin level; so we can say that patients with more severe infections tend to experience significant decreased in albumin level during acute phase of infection. Besides correlating the CRP with albumin level and comparing initial CRP level between the two groups with changes in albumin levels, the study also tried to measure the risk of albumin level reduction in patients with high initial CRP levels. The risk of albumin level reduction is higher in patients with higher initial CRP levels compared to patients with low initial CRP levels (RR = 2.12), but this was not statistically signi cant. However, a 2.12 times (or 112%) increase in risk could not just be ignored, because it is clinically important. There are two things that might contribute to why the nding was, as a forementioned, not statistically signi cant. These are also the limitations of this study. First, the observation time of albumin level reduction is too short (5 days) in this study. It might not be enough to observe changes in albumin level as negative acute-phase protein in ve days, because the half-life of albumin ranges between 19 to 21 days. 14,17 Therefore, this study uses a 10% cut-off point of albumin level reduction in ve days by predicting that the albumin level will reach a maximum 25% decrease during acute infection/in ammation in nine days. 8,16 However, on the other hand, longer observation on albumin level decrease might be biased by nutritional factors. Second, inappropriate sample size. A cohort study to look for risk differences for albumin level reduction in patients 16

5 Vol 39 Number 1 January - March 2007 with high and low CRP levels needs an estimated sample size of at least 45 patients. However, due to limitations in time and costs, also considering that the number of subjects was adequate to prove the main hypothesis (to look for correlations and mean differences) without lowering the power of the study, the study was terminated after 26 patients were included. As a consequence, in a cohort study to nd relative risk differences, this lack of sample affects the signi cancy of the result obtained. Questions are raised to whether nutritional factors i.e. previous nutritional status and protein diet during observation in uence the incidence of albumin level reduction in this study? Two hypothetical answers maybe should be consider. First, by looking at Table 2, the initial nutritional status (the Mini Nutritional Assessment [MNA] score, body mass index [BMI], and initial albumin level) are seem similar between the two groups, so the previous nutritional factors not in uence this reduction. Second, it is undoubtful that the rate of albumin synthesis is also affected by nutritional intake; 28 however, considering the 21-days half-life of albumin the intake contribution to albumin level reduction in ve days is very low. To prove these hypotheses, and also to correct the limitations of this study, we suggest to conduct a similar study with a larger sample size for a longer observation time, also controlling for nutritional status and protein intake as variables that could in uence albumin level during hospitalization. Considering that there were a lot of excluded conditions in this study to eliminate the interference of factors other than CAP to the CRP and albumin level, generalization of the study ndings to larger populations should be done cautiously. Application to populations outside of the study subjects or to larger population could be done as long as associated diseases or conditions in the elderly are taken into account. CONCLUSION From the result of this study, we could conclude that in hospitalized elderly patients with CAP there were (1) a correlation is found between initial CRP level with the percentage of albumin level decrease, and (2) a tendency that high initial CRP level will cause albumin level decrease during hospitalization; however, (3) risk differences of a reduced albumin level during hospitalization between patients with high CRP levels compared to low CRP levels has not yet been obtained. The two main results of this study show that a decrease of albumin level during the rst days of hospitalization in elderly patients with acute infection is associated with the severity of infection. C-reactive Protein Levels and Decrease of Albumin Levels ACKNOWLEDGEMENT The authors acknowledge and conveys gratitude to Suhardjono, M.D.,PhD, consultant nephrologisthypertensive disease, and Aru W. Sudoyo, M.D, PhD, consultant hematologist and medical oncologist for their guidance and supervision throughout the study. REFERENCES 1. Pinner RW, Teutsch SM, Simonsen L, Klug LA, Graber JM, Clarke MJ, dkk. Trends in infectious diseases mortality in the United States. JAMA. 1996;275: Lokakarya Survei Kesehatan Rumah Tangga, Cisarua. Badan Litbang Depkes RI Data Subbagian Rekam Medis RSUPN Dr. Cipto Mangunkusumo, Jakarta Data Subbagian Geriatri Bagian Ilmu Penyakit Dalam RSUPN Dr. Cipto Mangunkusumo Jakarta Meehan TP, Fine MJ, Krumholz HM, Scinto JD, Galusha DH, Mockalis JT, dkk. Quality of care, process, and outcomes in elderly patiens with pneumonia. JAMA. 1997;278: Koivula I, Sten M, Makela PH. Risk factors for pneumonia in the elderly. Am J Med. 1994;96: Nicholson JP, Wolmarans MR, Park GR. The role of albumin in critical illness. Br J Anaesth. 2000;85: Hedlund JU, Hansson LO, Ortqvist AB. Hypoalbuminemia in hospitalized patients with community-acquired pneumonia. Arch Intern Med. 1995;155: Sullivan DH, Sun S, Walls RC. Protein-energy undernutrition among elderly hospitalized patients. A prospective study. JAMA. 1999;281: Franch-Arcas G. The meaning of hypoalbuminaemia in clinical practice. Clin Nutr. 2001;20(3): Castell JV, Gomez-Lechon MJ, David M, Fara R, Trullenque R, Heinrich PC. Acute-phase response of human hepatocytes: regulation of acute-phase protein synthesis by interleukin-6. Hepatology. 1990;12: Moshage HJ, Janssen JAM, Franssen JH, Hafkenscheid JCM, Yap SH. Study of the molecular mechanism of decreased liver synthesis of albumin in inflammation. J Clin Invest. 1987;79: Ruot B, Breuille D, Rambourdin F, Bayle G, Capitan P, Obled C. Synthesis rate of plasma albumin is a good indicator of liver albumin synthesis in sepsis. Am J Physiol Endocrinol Metab. 2000;279:E Fleck A, Raines G, Hawker F, Trotter J, Wallace PI, Ledingham IM, Calman KC. Increased vascular permeability: a major cause of hypoalbuminaemia in disease and injury. Lancet. 1985;1: Ruot B, Papet I, Bechereau F, Denis P, Buf ere C, Gimonet J, dkk. Increased albumin plasma efflux contributes to hypoalbuminemia only during early phase of sepsis in rats. Am J Physiol Regul Integr Comp Physiol. 2003;284:R Gabay C, Kushner I. Acute-phase proteins and other systemic respons to in ammation. N Engl J Med. 1999;340: Pepys MB, Hirschfiels GM. C-reactive proten: a critical update. J Clin Invest. 2003;111: Smith RP, Lipworth BJ, Cree IA, Spiers EM, Winter JH. C-reactive protein: a clinical marker in community-acquired pneumonia. Chest. 1995;108: Almirall J, Bolibar I, Toran P, Pera G, Boquet X, Balanzo X, 17

6 Kuntjoro Harimurti, et al Sauca G, for The Community-Acquired Pneumonia Maresme Study Group. Contribution of C-reactive protein to the diagnosis and assessment of severity of community-acquired pneumonia. Chest. 2004;125: Samosir DRS. Kadar C-reactive protein sebelum dan setelah terapi antibiotik pada pasien usia lanjut dengan pneumonia komunitas yang dirawat di rumah sakit. Jakarta: Fakultas Kedokteran Universitas Indonesia, Tesis. 21. Kaysen GA, Stevenson FT, Depner TA. Determinants of albumin concentration in hemodialysis patients. Am J Kidney Dis. 1997;29: Kaysen GA, Greene T, Daugirdas JT, Kimmel PL, Schulman GW, Toto RD, dkk, and the HEMO Study Group. Longitudinal and cross-sectional effects of C-reactive protein, equilibrated normalized protein catabolic rate, and serum bicarbonate on creatinine and albumin levels in dialysis patients. Am J Kidney Dis. 2003;42: Kaysen GA, Dubin JA, Muller HG, Rosales LM, Levin NW, and the HEMO Study Group. The acute-phase response varies with time and predicts serum albumin levels in hemodialysis patients. Kidney Int. 2000;58: Acta Med Indones-Indones J Intern Med 24. Strandberg TE, Tilvis RS. C-reactive protein, cardiovascular risk factors, and mortality in a prospective study in the elderly. Arterioscler Thromb Vasc Biol. 2000;20: Dong Q, Wright JR. Expression of C-reactive protein by alveolar macrophage. J Immunol. 1996;156: Mold C, Rodic-Polic B, and Du Cloz TW. Protection from Streptococcus pneumoniae infection by C-reactive protein and natural antibody requires complement but not FcgReceptors. J Immunol. 2002;168: Kuczmarski MF. Nutritional status of older adults. Nutrition and aging. Edisi ke-2. Missouri: Mosby-Tear Book Inc; p Kaysen GA, Jones H, Martin V, Hutchison FN. A low-protein diet restricts albumin synthesis in nephritic rats. J Clin Invest. 1989;83: Long CL, Jeevanandam M, Kim BM, Kinney JM. Whole body protein synthesis and catabolism in septic man. Am J Clin Nutr. 1977;30(8):