| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From The Clinical Research Centers |
Clinical Endocrinology Branch, National Institute of Diabetes, Digestive, and Kidney Diseases (N.J.S.); Pediatric Oncology Branch, Division of Clinical Sciences, National Cancer Institute (S.J.C.); and Developmental Endocrinology Branch, National Institute of Child Health and Human Development (L.K.N.), National Institutes of Health, Bethesda, Maryland 20892
Address correspondence and requests for reprints to: Lynnette K. Nieman, M.D., Clinical Director, National Institute of Child Health and Human Development, National Institutes of Health, Building 10, Room 9S261, 10 Center Drive, MSC 1834, Bethesda, Maryland 20892-1834. E-mail: niemanl{at}nih.gov
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Charts of 54 consecutive patients with untreated EA, without underlying diagnosis of small cell carcinoma of the lung, were reviewed, and types of infections, white blood cell (WBC) count, fever, as well as the glucocorticoid indices [0800 h F, daily urine F excretion (UFC), and daily urine 17-hydroxysteroid/g creatinine excretion (17OHS)], were recorded.
Thirty-five patients had no or clinically mild infection; the remaining 19 patients had severe, systemic infection (n = 13) and/or sepsis (n = 6), including either bacterial or opportunistic pathogens or both (73.7%, 42.1%, and 13.8%, respectively). The latter group of patients had significantly higher indices of hypercortisolism (F, UFC, and 17OHS) than those with mild or no infections, but these indices did not correlate with temperature or WBC count. Thresholds for identifying severe infection were selected for maximal positive predictive value and were: F, 43.1 µg/dL; UFC, 2000 µg/day; and 17OHS, 35 mg/g creatinine. The most accurate discriminator for severe infection was 17OHS, based on a positive predictive value of 64.7%.
Our data strongly suggests that the likelihood for a bacterial or opportunistic infection in CS patients, even without underlying small cell carcinoma of the lung, is greatest in patients with extreme hypercortisolism. The predictive value of total WBC count or the presence of an elevated temperature is not sufficient to identify patients with severe, life-threatening infection.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
High circulating levels of either exogenous or endogenous glucocorticoids impair immune function and predispose to infectious complications (9, 10, 11, 12, 13, 14, 15). The major effect of glucocorticoids is primarily on cell-mediated immunity (16, 17, 18, 19). Although it seems logical to assume that indices of cortisol (F) production could accurately predict infections in CS of different etiologies [not including small cell carcinoma of the lung (SCCL)], this has not been demonstrated firmly. Furthermore, the clinical profile of side effects differs between hypercortisolism from an exogenous source (usually iatrogenic) compared with an endogenous source (12, 20, 21). A meta-analysis of 71 trials involving more than 2000 glucocorticoid-treated patients with nonmalignant pathology showed that the relative risk for infection was 1.5 times that of controls (12). The degree of risk for infection was a complex function of the level and duration of glucocorticoid excess, the dosing schedule, clearance of glucocorticoids, the presence of coexisting immunosuppressive therapy, and the nature of the underlying host disease. Exogenous glucocorticoid use poses markedly different risks of infection in different disease states. This should be viewed in contradistinction to the relatively limited range of pathological processes causing endogenous hypercortisolism, where perhaps the relationship between the level of glucocorticoid excess and the risk of infection could be more easily assessed.
Previous reports of infectious complications in EA-related CS, most caused by SCCL, suggest that the frequency and severity of infectious complications is directly proportional to the degree of hypercortisolemia (4, 5, 6, 7, 8). These studies emphasized the high prevalence of unusual infections in these patients with opportunistic pathogens, including Candida, Cryptococcus, Aspergillus, Nocardia, Pneumocystis carinii, Toxoplasma, Cryptosporidium, the herpes viruses (cytomegalovirus, herpes simplex, and varicella-zoster), Mycobacterium tuberculosis, Mycobacterium avium-intracellulare, and Listeria monocytogenes. This literature is clouded by the inclusion of a significant number of patients with underlying SCCL, a fact that is of importance for several reasons because these patients have tumor-related features that could confound the observed associations. In patients with EA-related CS due to SCCL, treatment of the underlying hypercortisolism with either surgical or medical adrenalectomy prior to the initiation of cytoreductive chemotherapy reduces the frequency of infections after the initiation of chemotherapy (6, 8). Furthermore, a case report has described reasonable response of disseminated fungal infection with Alternaria spp. to adrenolytic therapy in a patient with EA-related CS (22).
This study considered the significance of extremely high levels of endogenous F production and their possible correlation to a higher incidence and/or severity of infectious complications in patients with EA-related CS. To evaluate this, we examined retrospectively patients with EA-related CS and extreme hypercortisolemia who did not have a diagnosis of SCCL and established a firm association between the severity of infectious complications and indices of F production. Patients with SCCL have been excluded from our analysis because of the possible role of the underlying malignancy and therapy administered (which can be immunosuppressive) (15). Because exogenous glucocorticoids suppress fever and elevate white blood cell (WBC) counts, we also sought to determine whether the cortisolemic indices might be better predictors of severe infections in our patients than these "classical" features of infection.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The charts of 54 consecutive patients with untreated EA-related CS who were admitted to the National Institutes of Health Endocrinology Service between January 1986 and January 1996 were reviewed. All patients participated in clinical protocols approved by the institutional review board of the National Institute of Child Health and Human Development. Informed consent was obtained from all patients at the time of their initial evaluation with regard to future use of clinical data for retrospective studies. The diagnosis of CS due to EA was assigned based on either the histology of the biopsied tumor (including the demonstration of immunopositivity for ACTH) or the biochemical results from bilateral inferior petrosal sinus sampling (23) and lack of dynamic responses to dexamethasone suppression (24, 25) and/or CRH stimulation tests (26).
To evaluate the ability of outcome measures to predict severe infections, patients were divided into group A (without evidence of infection or with mild infection) and group B (with severe infection, with or without sepsis). Infections were classified according to the documented pathogen (bacterial, protozoal, viral, fungal) and also on the basis of clinical severity. Severe infections were identified on the basis of one of the following: formation of an abscess, isolation of a pathogen from blood cultures, or prolonged continuous administration of iv antibiotics after failure of the oral or im route to control the infection. All other infections were classified as "mild."
With regard to clinical diagnostic features of mild infection, the majority of patients were asymptomatic, except for those with fungal skin infections. In these patients, the diagnosis of infection was made incidentally during evaluation of CS or by an incidental abnormal laboratory finding (i.e. bacteriuria). With regard to clinical diagnostic features of severe infection, patients presented mainly with vague "ill" complaints or poorly localized pain. Patients with sepsis presented with hypotension, especially in the setting of postoperative sepsis. Unfortunately, due to the retrospective nature of this study, it was impossible to document the initial clinical manifestations of infection(s) in all patients, and, hence, no correlations were attempted between the presence (or absence) and severity of such clinical features and level of hypercortisolemia. However, it should be noted that no "classical" features of infections (i.e. localized pain, erythema, edema, fever, or chills) were present in the majority of patients with severe infections for which clinical data existed. Furthermore, the duration of exposure to high levels of endogenous glucocorticoids was impossible to assess (as is the case in most patients with CS at their initial presentation).
Data were collected on the underlying causes of the EA-related CS, infectious complications, highest total WBC count at presentation, highest oral temperature, and highest recorded values of glucocorticoid indices [0800 h F, daily urine F excretion (UFC), and daily urine 17-hydroxysteroid/g creatinine (Cr) excretion (17OHS)] before identification of an infection (where present). Parameters were collected from a single hospitalization during which the patients were hypercortisolemic. If patients were admitted more than once, the only parameters taken into consideration were the ones related to the admission during which an infectious complication developed.
Data analysis
Clinical features, biochemical parameters, WBC count, and temperature were analyzed to determine whether they were associated with the development of an infectious complication. The Mann-Whitney U test for continuous variables and Fisher’s exact test for categorical variables were used for comparisons between groups A and B. Pearson linear correlation coefficients (r) were calculated between pairs of all variables tested (F, UFC, 17OHS, WBC count, and temperature). Data are expressed as mean ± SEM. A two-tailed P value of <0.05 was considered significant. Bivariate plots were constructed for all variables separated by group (A vs. B), and cut-off points were calculated by box-plot analysis, taking into consideration any outliers. The cut-off points were selected for achieving the maximum positive predictive value (PPV) for comparisons between groups A and B. Once a cut-off point was identified for maximal PPVs, sensitivity and specificity were subsequently calculated for that cut-off point. The StatView 5.0 statistical analysis program (1998) was used for all analyses (StatView; SAS Institute, Inc., Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The average age of the 54 patients was 36.1 ± 16.1 yr (range, 8–71 yr). Twenty patients were male (37.3%), and 34 patients were female (62.7%). There were 45 Caucasian (83.3%), 3 African American (5.6%), and 6 Hispanic (11.1%) patients.
The causes of EA-related CS are shown in Table 1
and included carcinoid tumors,
neuroendocrine neoplasms (other than SCCL), and EA of unknown
(anatomically nondefined) origin. There were no cases of clinically
overt SCCL among the patients studied.
|
Thirty-five patients were identified in group A, with either no evidence of infection (n = 29) or mild infection (n = 6). Group B (n = 19) included patients with severe infection without sepsis (n = 13) or sepsis (n = 6).
The distribution of the types of infectious complications in the study
population is shown in Table 2. Notably,
infections in patients from group B were divided between bacterial
(73.7%) and opportunistic pathogens (42.1%); 13.8% of patients in
this group had more than one pathogen identified in either one or
different sites, and, hence, their infection was considered to be
"mixed."
|
All cortisolemic indices were significantly higher in group B
vs. group A, including serum F (63.7 ± 7.2
vs. 41.2 ± 4.7 µg/dL, P < 0.03)
[normal (nl) values (National Institutes of Health laboratory), 7–20
µg/dL], UFC (6517 ± 2454 vs. 1858 ± 394
µg/day, P < 0.004) (nl values, <90 µg/day), and
17OHS (45.7 ± 7.8 vs. 23.2 ± 3.5 mg/g Cr,
P < 0.002) (nl values, 2.0–6.8 mg/g Cr) (Fig. 1). Correlations between pairs of variables
for all three parameters (F, UFC, and 17OHS) in both groups were highly
significant (r = 0.30 to 0.86, P < 0.0001 to
0.015). There was no significant correlation of these indices with
temperature, WBC count, or other clinical features (patient’s age and
type, severity, or site of infection) (data not shown).
|
. The highest PPV for comparisons
between groups A and B was observed for 17OHS at a level of 35 mg/g Cr.
Since there were no cut-off points with complete discrimination between
the groups, each cut-off point can be considered as a reasonable
predictor for the development of severe infectious complications, with
the risk increasing as the levels of cortisolemic indices increase
further. The scatter of values for plasma F and UFC in group A was
considerably greater than that seen for 17OHS in the same group, thus
weakening the significance of the observed thresholds between the two
groups. On the other hand, although the 17OHS parameter was the one
most strongly associated with the development of severe infection, it
is no longer considered a "routine" measure of hypercortisolism in
clinical practice, with UFC being more widely used. Indeed, no patient
with a UFC <735 µg/day developed a severe infection (Fig. 1B
, arrow); hence, patients below this UFC cut-off point seem
less likely to develop such a complication.
|
The distribution of total WBC count and body temperature in the
patient groups is shown in Fig. 2. No
reliable and clinically relevant cut-off points are evident. Notably,
only 5 of 19 patients (26.3%) with severe infections had oral
temperatures 
38 C, and the highest temperature noted in that group
was 38.6 C. These indices did not correlate with F, UFC, or 17OHS (as
mentioned above) or other clinical features (patient’s age and type,
severity, or site of infection). Although a left-shift in the
differential WBC count has been associated with severe infection, it
may also be associated with high levels of endogenous or exogenous
glucocorticoids. Viral infections may induce lymphocytosis, but this
was not analyzed in our cohort because it was not consistently tested
or recorded.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The data from this study emphasize that both bacterial and opportunistic infections occur in ill hypercortisolemic patients, sometimes within the time span of a single hospitalization. These findings are similar to another recently reported series of patients with PCP who were not infected with human immunodeficiency virus (37), the majority of whom received either high doses of exogenous glucocorticoids or were exposed to lower doses for a prolonged period of time. In that study, a wide variety of concurrent nonprotozoal organisms were also isolated, including bacterial pathogens (37). Similarly, Bakker et al. (38) have described a case of CS with severe F excess that was complicated by simultaneous infections with multiple opportunistic pathogens, as well as Staphylococcus aureus sepsis. In our study, it is notable that the majority of events in group B (24 of 29 cases or 82.8%) were also related to bacterial pathogens. The effect of glucocorticoids on the response to bacterial pathogens is not as well defined as it is for classical, cell-mediated pathways that are critical for antifungal and anti-PCP immunity. However, it seems that high levels of F interfere with clearance of bacteria, as indicated by the number of events associated with aberrant host defense systems, such as abscess, pneumonia, and osteomyelitis.
The manifestations of infection are altered by the potent anti-inflammatory actions of glucocorticoids. Patients with CS have a marked inability to localize infection, as judged by either radiographic or clinical parameters (1), which in turn often leads to a significant delay in diagnosis and initiation of therapy. As a result, morbidity and mortality are increased (14, 39). We have shown that high levels of endogenous glucocorticoids, and not WBC counts or temperature, are reliable indicators for predicting severe infections in patients with untreated EA-related CS.
Our data add further evidence that sustained hypercortisolism in the setting of EA impairs of crucial host defense mechanisms in humans in a dose-dependent manner. In fact, initiation of prophylaxis for Pneumocystis carinii has been suggested in patients with CS when plasma F exceeds 90 µg/dL (38). The increased incidence and severity of infectious complications with progressively more elevated levels of F, UFC, and 17OHS suggest a significant role for increased and continuous clinical surveillance of patients with similarly severe hypercortisolism. On the other hand, several studies have shown that PCP can present in patients with CS after either surgical or medical treatment has been instituted to lower F levels (40, 41, 42). Pneumonia in these patients seems to become clinically manifest when the anti-inflammatory effects of hypercortisolemia are attenuated. This observation has led to the initiation of trimethoprim-sulfamethoxasole therapy in a prophylactic fashion in patients who have severe hypercortisolism before its definitive treatment (43). Accordingly, it is our current practice to initiate PCP prophylaxis in patients with elevated UFC levels at greatest risk before definitive management of their hypercortisolism.
The inability of measures of hypercortisolism to fully predict risk of infection (maximal PPV for thresholds, 64.7%) was, perhaps, counter-intuitive to the hypothesis that glucocorticoids per se impede the immune response. It is possible that in addition to the magnitude of exposure to high levels of glucocorticoids, the duration of such an exposure also modulates the host response to infection. Unfortunately, it was not possible to estimate accurately the duration of hypercortisolism before the development of infectious complications from the information in the patients’ records. Alternatively, it is possible that the genetic vulnerability to infection is unmasked differentially at varying levels of glucocorticoid exposure among different patients. If the latter hypothesis stands true, this could have important implications for the management of patients with not only endogenous hypercortisolism, but also exogenous (iatrogenic) hypercortisolism, a much larger population at risk for infection. Although our study is based on a fairly uniform patient population (i.e. patients with EA-related CS not due to SCCL), one would predict that our findings are applicable to other patients with profound hypercortisolism, whether the etiology is due to ACTH-dependent (pituitary or nonpituitary) vs. ACTH-independent CS. Indeed, patients with all forms of CS are susceptible to infection with opportunistic pathogens. Our study was restricted to a select subgroup of patients to minimize the contribution of complicating and potentially confounding factors, such as antineoplastic therapy for SCCL.
In conclusion, our study demonstrates that although susceptibility to infection is likely to be present in patients with very high F production indices, the actual basis for the development of infection is multifactorial. Since our data were analyzed in a retrospective fashion, we have only partially resolved the obvious obstacle of making assumptions about the risk of severe infection in patients with CS. Thus, we strongly believe that the importance of our findings indicates the need for prospective observation of hypercortisolemic patients at risk of infectious complications.
|
Footnotes |
|---|
Received August 2, 1999.
Revised September 27, 1999.
Accepted September 30, 1999.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
B activity
through induction of IB synthesis. Science. 270:286–290.This article has been cited by other articles:
 |
|
 |
|
  L. K. Nieman, B. M. K. Biller, J. W. Findling, J. Newell-Price, M. O. Savage, P. M. Stewart, and V. M. Montori The Diagnosis of Cushing's Syndrome: An Endocrine Society Clinical Practice Guideline J. Clin. Endocrinol. Metab., May 1, 2008; 93(5): 1526 - 1540. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Isidori, G. A. Kaltsas, C. Pozza, V. Frajese, J. Newell-Price, R. H. Reznek, P. J. Jenkins, J. P. Monson, A. B. Grossman, and G. M. Besser The Ectopic Adrenocorticotropin Syndrome: Clinical Features, Diagnosis, Management, and Long-Term Follow-Up J. Clin. Endocrinol. Metab., February 1, 2006; 91(2): 371 - 377. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Ilias, D. J. Torpy, K. Pacak, N. Mullen, R. A. Wesley, and L. K. Nieman Cushing's Syndrome Due to Ectopic Corticotropin Secretion: Twenty Years' Experience at the National Institutes of Health J. Clin. Endocrinol. Metab., August 1, 2005; 90(8): 4955 - 4962. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Mirani, I. Elenkov, S. Volpi, N. Hiroi, G. P. Chrousos, and T. Kino HIV-1 Protein Vpr Suppresses IL-12 Production from Human Monocytes by Enhancing Glucocorticoid Action: Potential Implications of Vpr Coactivator Activity for the Innate and Cellular Immunity Deficits Observed in HIV-1 Infection J. Immunol., December 1, 2002; 169(11): 6361 - 6368. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Krakoff, C. A. Koch, K. A. Calis, R. H. Alexander, and L. K. Nieman Use of a Parenteral Propylene Glycol-Containing Etomidate Preparation for the Long-Term Management of Ectopic Cushing's Syndrome J. Clin. Endocrinol. Metab., September 1, 2001; 86(9): 4104 - 4108. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Vassilopoulou-Sellin Paraneoplastic Hypercortisolism as a Risk Factor for Severe Infections in Patients with Malignant Diseases J. Clin. Endocrinol. Metab., February 1, 2001; 86(2): 947a - 948. [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
