The acute phase response (APR)⁵ is a predictable, stereotyped metabolic response to infection and injury. It is stimulus nonspecific and is characterized by leukocytosis, fever, increased synthesis of acute phase plasma proteins, increased plasma copper concentrations, increased resting energy expenditure and depressed plasma iron and zinc concentrations (Beisel 1977). This response is mediated by a number of cytokines, including tumor necrosis factor, interleukin 1 (IL-1) and interleukin 6 (IL-6). Many of these cytokine-mediated alterations, including fever, hypoferremia, leukocytosis and elaboration of acute phase proteins, beneficial to the host.

Zinc is involved in the elaboration of several cytokines. The synthesis of interferon (Reardon and Lucus 1987) and leukocyte migration inhibitory factor are zinc dependent (Bendtzen 1980, Bendtzen and Mayland 1982). Zinc influences in vitro human leukocyte production of the cytokines tumor necrosis factor, IL-1 and IL-6 (Scuderi 1990). The regulation of leukocyte cytokine production by zinc suggests that this trace metal may play an important role in controlling the APR. During the APR, serum concentrations of zinc decline 10 to 69% because of hepatic sequestration by metallothionein (Schroeder and Cousins 1990). The decline in plasma zinc may serve as an important regulator of leukocyte cytokine production and ultimately influence or control the APR.

Despite the lack of knowledge concerning zinc sequestration during the APR, the importance of zinc for wound healing and in immunity has prompted some clinicians to routinely supplement parenteral solutions with additional zinc for catabolic, critically ill patients (Freund 1986, Weinsier and Morgan 1993). Current guidelines for parenteral trace elements recommend supplementing the standard daily dose of zinc (5 mg elemental zinc) with 2-4 mg ZnSO₄/d for catabolic patients (Shils et al. 1979). If the sequestration of zinc represents a host mechanism to control leukocyte cytokine production, prevention of the decline in serum zinc by use of supplemental parenteral zinc could be detrimental to the host.

The goal of this study was to investigate the effect of zinc supplementation on the APR. The specific aim was to determine whether parenteral zinc supplementation during the APR maintains serum zinc concentrations within the "normal" range and results in an exaggerated APR, as evidenced by hyperthermia and elevated IL-6 and ceruloplasmin levels.

METHODS

Patients were excluded if they had conditions known to influence zinc, copper or immune status. Exclusion criteria included the presence of protein energy malnutrition, current pregnancy or lactation, acrodermatitis enteropathica, prophyria, chronic blood loss, hepatic disease, sickle cell anemia, current parasitic infection, cancer, HIV infection, steroid therapy, current alcoholism, diabetes mellitus, renal failure, Wilson's disease, malabsorption disorders including celiac disease and inflammatory bowel disease (acceptable if on home TPN), and prior use (within 6 mo) of routine supplemental trace elements at doses greater than the U.S. Recommended Dietary Allowances. Informed consent was obtained from all participants prior to study enrollment. The study protocol was approved by the Ethics Committee of the Institutional Review Board of the University of Michigan Medical Center.

The number of subjects required per group was calculated using the normal approximation for two independent means of equal sample size. Power curves reflecting sample sizes and rates of clinically detected differences were determined for IL-6 levels. On the basis of a mean IL-6 concentration of 0.085 ± 0.01 U/L) (Lesser et al. 1991), the number of participants required to detect a 20% difference with a significance level of 0.05 and a power of 0.80 was estimated to be 22 per group.

Block randomization was used for individual assignment of 0 or 30 mg elemental zinc sulfate per day for the first 3 d of TPN administration. The participants' TPN solutions were compounded by technicians in the Department of Pharmacy. Table 1 provides a complete description of the standard multivitamins and trace elements included in the TPN solutions of all participants. No participants received any iron in their TPN solutions while enrolled in the study. All participants returned to the standard zinc therapy of 5 mg/d at the end of the third day of TPN.

Table 1. Composition of vitamins and trace elements included in all participants' parenteral nutrition solutions [View Table]

Baseline serum samples were obtained from participants upon entry into the study (prior to receiving any TPN) and on d 1, 2 and 3 following study enrollment. Following a 1:7 dilution, serum zinc concentrations were determined by atomic absorption spectrophotometry (Thermo Jarrell Ash, model Smith-Hieftje 4000, Franklin MA). Interleukin 6 concentrations were assessed using the methods of Aarden et al. (1987) for the detection of hybridoma growth factor. Ceruloplasmin was measured using techniques described by Lehmann et al. (1974). Serum copper concentrations were measured using flame atomic absorption spectrophotometry (Hill and Miller 1983). Urine collections (24 h) were obtained on d 1, 2 and 3 following study enrollment. Urinary zinc concentrations were measured by atomic absorption spectrophotometry. Temperatures were obtained from the patients' bedside charts (Diatek thermometers, Welch Allyn Co., San Diego, CA). The temperature reported just prior to the time of enrollment in the study and the highest temperatures on d 1, 2, and 3 of TPN administration were recorded. The highest temperatures rather than average temperatures were used to reflect the magnitude of the APR. We could not control the administration of antipyretic medications. As a standard of care, hospitalized patients are given antipyretics in various amounts and time intervals with presentation of fever. An average temperature would reflect both the treated and untreated fever whereas the highest temperature would better represent the true magnitude of the APR.

To ascertain whether randomization was effective in distributing possible confounding characteristics between the zinc-supplemented and unsupplemented groups, chi-square tests were used to assess the homogeneity of categorical variables (sex, disease status) A nonparametric approach (Wilcoxon) (Cody and Smith 1991) was used to assess effective randomization between study groups for the continuous variables (age, weight, temperature, serum IL-6, zinc, copper and ceruloplasm). To determine whether zinc supplementation had been administered, serum zinc and urinary zinc concentrations of the two study groups were compared. To test whether the observations across time (baseline and d 1-3 TPN) in treated vs. untreated were characterized by similar slopes and intercepts, serum zinc, copper, ceruloplasmin, IL-6 and temperature were compared using repeated measures ANOVA (SAS/STAT version 6, SAS Institute, Cary, NC).

RESULTS

Table 2. Baseline characteristics of study groups following randomization for supplemental parenteral zinc1 [View Table]

As would be expected with successful zinc supplementation, there was a significant difference (P = 0.009) in serum zinc concentration between the study groups on d 1, 2 and 3 of TPN (Table 3). Due to an interaction in this variable we were unable to use repeated measures ANOVA to assess an across-study interval difference. The serum zinc concentrations of the controls remained virtually unchanged, whereas the concentrations of the supplemented group increased by approximately 50% during the first 3 d of TPN. There was a significant difference in urine (P = 0.0004) zinc concentrations between the two groups over the 3-d observation period. The concentration of urinary zinc in the control group decreased from 9.79 to 6.12 µmol/L over the three study days. This represents a 37% decline and is characteristic of what would be expected in patients without supplementation.

Table 3. Serum and urine zinc concentrations in controls and zinc supplemented groups receiving total parenteral nutrition (TPN) during the acute phase response1 [View Table]

There was a significant difference in temperature between the treated and untreated groups (P = 0.035) over the 4-d study period (Fig. 1). Individual P values for differences at the four time intervals were as follows: baseline, P = 0.48; d 1 TPN, P = 0.06; d 2 TPN, P = 0.25; d 3 TPN, P = 0.01. Temperatures of participants in the zinc-supplemented group became significantly more elevated over the 3 d of zinc supplementation during the APR than did temperatures of the unsupplemented group. No significant differences between study groups were observed for serum IL-6, ceruloplasmin or copper concentration (data not shown).

A subgroup analysis was undertaken that removed from analysis those participants with serum zinc concentrations greater than normal (serum zinc 19.98 µmol/L). This was done to examine the possibility that the amount of zinc administered resulted in serum zinc concentrations that exceeded normal concentrations and in turn induced an exaggerated APR. Despite the loss of power that occurred with the removal of seven participants (33% of the original supplemented group), the differences in temperature across the time intervals remained significant (P = 0.036, Table 4). The individual P values for differences between the two groups at the four time intervals were as follows: P = 0.90 at baseline, P = 0.08 on d 1 of TPN, P = 0.11 on d 2 of TPN, P = 0.029 on d 3 of TPN.

Table 4. Temperatures from baseline to d 3 total parenteral nutrition (TPN) in a subgroup of zinc-supplemented and unsupplemented patients during the acute phase response1,2 [View Table]

As stated, IL-6 concentrations were not significantly different between the study groups. To determine whether an association between IL-6 concentrations and change in temperature existed, a correlation analysis was conducted. The correlation between fever and plasma IL-6 activity for all participants was r = 0.55 (P = 0.0002) on d 3 of TPN. When separated by treatment groups, the correlations between temperature and IL-6 concentrations were r = 0.46 (P = 0.04) for supplemented patients and r = 0.57 (P = 0.007) for those that did not receive zinc on d 3 of the study.

DISCUSSION

Our findings of a more pronounced febrile response in humans supplemented with zinc during an APR are consistent with the very limited animal literature. Alterations in the febrile response in rabbits following zinc administration during the APR were reported in 1978 by Mapes et al. (1978). The animals were supplemented with intravenous zinc chloride (1.6 mg/kg body wt) 1 h before intraperitoneal IL-1 infusions. Control rabbits that did not receive zinc before the IL-1 challenge had a monophasic fever that peaked 45 min after IL-1 administration and returned to basal levels within 70-80 min. When zinc was administered 1 h before the IL-1 challenge, fevers were approximately 40% higher than those obtained in control rabbits. The fevers in the zinc-treated animals lasted for 4-6 h before returning to basal levels. When a minimal amount of zinc chloride (0.04 µ/kg body wt) was given simultaneously with the IL-1, fevers of prolonged duration occurred, but the level of the fever did not differ from that in control animals. Results of this investigation preceded the panel of experts guidelines for intravenous trace element supplementation (Shils et al. 1979), which recommends 5 mg zinc/d as a standard dose in TPN patients, with an additional 2-4 mg/d for catabolic patients. This continues to be the standard amount of zinc used in parenteral nutrition solutions today.

There are several reports of alterations in thermoregulation in zinc-deficient animals. Topping et al. (1981) found zinc-deficient rats had lower rectal temperatures than controls. When those animals were exposed to the cold, they were more prone to developing hypothermia than were zinc-adequate animals. More recently, Cossack (1991) reported alterations in the febrile response in zinc-deficient rats following the injection of E. coli. The zinc-deficient animals had a significantly lower febrile response 5-8 h after the endotoxin injection than did the pair-fed controls. Of note is the significant (P < 0.001) weight loss that occurred in the zinc-deprived rats during the 6-wk study period. At the end of the 6-wk study period average body weights were 142.2 ± 24.2 and 182.8 ± 7.7 g for zinc-deficient rats and pair-fed controls, respectively. Typically, animals fed zinc-deficient diets have lower weight gain than pair-fed controls (O'Dell and Reeves 1989). Many species respond to zinc deficiency with anorexia and cyclic feeding (Clegg et al. 1989). It is thought that the reduced food intake induces muscle catabolism and release of stored muscle zinc into the plasma. This zinc is then available for use by the liver and other tissues for zinc-dependent processes. Whether zinc deficiency per se or general protein energy malnutrition was responsible for the responses observed in Cossack's study cannot be determined.

Alterations in thermoregulation due to zinc deficiency have not been reported in humans. However, it has been known for many years that uncomplicated starvation in humans reduces core temperatures (Golden 1988).

In addition to temperature, serum IL-6 activity was determined as a measure of APR severity. Because of the great variability, no difference in IL-6 concentration was demonstrated between the control and zinc-supplemented groups. Numerous studies have observed associations between severity of illness or inflammatory process with IL-6 concentrations (Damas et al. 1992, Lesser et al. 1991, Nijsten et al. 1987, Viedma et al. 1992). Only Nijsten et al. (1987) reported a significant (r = 0.61, P < 0.001) correlation between IL-6 concentrations and body temperatures. In our study, correlations between temperature and IL-6 concentrations were significant in both the supplemented and control groups. To our knowledge, ours is the first study with patients experiencing a mild APR to demonstrate significant correlations between serum IL-6 concentrations and temperatures. Collectively, the significantly higher temperatures in the zinc-supplemented group and the correlation with IL-6 concentrations further corroborate the hypothesis that zinc supplementation during the APR results in a more exaggerated response.

Zinc supplementation induces alterations in copper status potentially by a competitive interaction between zinc and copper for intestinal absorption. Zinc intake stimulates enterocyte metallothionein synthesis (Richards and Cousins 1975). Metallothionein sequesters copper much more avidly than zinc, making the copper unavailable for transfer and reducing copper absorption. Oral zinc supplements as low as 18.5 mg/d have been reported to induce copper imbalances (Festa et al. 1985). During the APR, endotoxin, cytokines and glucocorticoid hormones have all been demonstrated to increase metallothionein mRNA in the thymus, liver and bone marrow (Hambidge et al. 1986). This induction is associated with a redistribution of zinc (Cousins 1985, Cousins and Leinart 1988). The avid binding of copper by metallothionein, the enhanced production of this protein during zinc administration, and stress may lead to alterations in serum copper concentrations.

The gastrointestinal homeostasis of zinc absorption via induction of metallothionein was avoided in our study through the use of parenterally delivered zinc. Several weeks of oral zinc supplementation are typically required to induce reductions in copper and ceruloplasmin concentrations (Prasad et al. 1978). As would be expected, we did not observe any acute differences in serum copper or ceruloplasmin concentrations with our short-term parenteral zinc supplementation.

The use of zinc supplementation is expanding. In general, zinc is considered to be nontoxic and humans seem to tolerate fairly high intakes. The role of zinc in wound healing and in the immune system and the occurrence of zinc deficiency in patients with malabsorptive disorders, AIDS and diabetes are well established (King and Keen 1994). Oral zinc supplements are routinely recommended by physicians for patients with any of these conditions, presumably to enhance immune function.

The results from our study indicate that zinc supplements given parenterally to patients during a mild APR will result in an exaggerated APR. The effect of oral zinc supplementation on temperature during the APR remains unknown. Hempe et al. (1991) documented in rats that zinc absorption is not impaired during the APR; thus, oral zinc supplements may have the same impact on the APR as we observed. In some situations, such as in treatment of AIDS patients, the exaggerated response may prove to be beneficial. In other situations, in which reduced inflammatory response is desired, such as in treatment of patients experiencing a flare-up of Crohn's disease, ulcerative colitis or arthritis, an exaggerated APR may be detrimental. Future studies of these disease are warranted to delineate benefits achieved as well as to avoid any untoward effects that may occur from overzealous supplementation with this nutrient.

The effect of zinc supplementation during a severe APR has not been determined. Patients who may receive supplemental zinc include postoperative, trauma, burn and general intensive care unit patients. Many of the new enteral products that are marketed specifically for these patients contain more than 24 mg zinc/4180 kJ. It is not uncommon for these patients to require 12,540 kJ/d and thus receive more than 72 mg zinc/d. Further clarification of the role of zinc sequestration in severe APR is needed prior to empiric supplementation in this population.

We have demonstrated in a prospective, randomized, double-blinded clinical trial that supplementation of 30 mg of parenteral zinc for 3 d in patients experiencing a mild APR resulted in a significantly higher febrile response. This level of zinc supplementation did not result in differences in serum copper and ceruloplasmin concentrations between study groups. We also found temperatures to be significantly correlated with serum IL-6 concentrations.