Risk of Lymphoma Following Exposure to Calcineurin Inhibitors and Topical Steroids in Patients with Atopic Dermatitis

Systemic use of immunosuppressant agents increases the risk of lymphoma in transplantation. We performed a nested case–control study in the PharMetrics database to evaluate the association between topical immunosuppressants and lymphoma in a cohort of patients with atopic dermatitis. We identified cases of lymphoma and randomly selected four controls for each case, matched by length of follow-up. We used conditional logistic regression to calculate odds ratio (OR) and 95% confidence intervals (CIs) of the association between topical immunosuppressants and lymphoma. Two hundred and ninety-four cases of lymphoma occurred in 293,253 patients, 81 in patients younger than 20 years. The adjusted analysis yielded the following OR (95%CI) for: severity (OR 2.4; 95% CI 1.5–3.8), oral steroids 1.5 (1.0–2.4), ‘‘super potent’’ topical steroids 1.2 (0.8–1.8) , ‘‘low potency’’ topical steroids OR 1.1 (0.7–1.6); pimecrolimus 0.8(0.4–1.6), tacrolimus OR 0.8 (0.4–1.7), and concomitant topical steroids, pimecrolimus, and tacrolimus 1.0 (0.3–4.1). We did not find an increased risk of lymphoma in patients treated with topical calcineurin inhibitors. It is difficult to disentangle the effects of severity of disease on outcome versus the true effects of drugs. However, in the adjusted analysis, severity of AD was the main factor associated with an increased risk of lymphoma.

Journal of Investigative Dermatology (2007) 127, 808–816. doi:10.1038/sj.jid.5700622; published online 9 November 2006

INTRODUCTION

The prevalence of atopic dermatitis (AD) in Western countries has increased over the last 30 years (Fleming et al., 2001). AD affects approximately 20% of children and 1–3% of adults in developed countries (Williams et al., 1999). Patients suffer from itch, loss of sleep, bleeding from the skin and skin infection, and the disease also influences patients’ (1Risk Management Resources, 726 Route 202 South, Suite 320-361, Bridgewater, New Jersey, USA; 2St Vincent’s Clinical School, Faculty of Medicine, UNSW, Sydney, Australia; 3Risk Management Resources – EU, Zaragoza, Spain; 4Clinical Safety and Epidemiology, Novartis Farmece´utica. Gran V´ıa de las Cortes Catalanas, Barcelona, Spain; Department of Dermatology, Hopital Universitaire Purpan, Place du Dr Baylac, Toulouse Cedex, France and 6Clinical Research Dermatology, Novartis Pharma, Basel, Switzerland Correspondence: Dr Felix M. Arellano, Risk Management Resources, 726 Route 202 South, Suite 320-361, Bridgewater New Jersey 08807, USA. E-mail: [email protected] Abbreviations:   AD, atopic dermatitis; CI, confidence interval; MF, mycosis fungoides; NHL, non-Hodgkin lymphoma; OR, odds ratio; OTC, over the counter; RR, relative risk Received 16 April 2006; revised 21 September 2006; accepted 22 September 2006; published online 9 November 2006 ) families (Su et al., 1997). AD is primarily a childhood disease but persists through adulthood in up to 50% of cases (Williams, 1997). For many patients with AD, intermittent treatment with topical immunomodulatory drugs, that is, topical corticosteroids or topical calcineurin inhibitors, is necessary for disease control.

In the USA the incidence of lymphoma ranges from around one case per 100,000 persons in children aged 0–19 years to 54 cases per 100,000 in patients over age 50 (Surveillance, Epidemiology, and End Results (SEER) Program. http://www.seer.cancer.gov. Last accessed 16 April 2006). Systemic use of immunosuppressive agents increases the risk of lymphoma, especially after transplantation (Opelz and Henderson, 1993). The risk of lymphoma in patients under- going organ transplantation is closely related to the intensity of immunosuppression (i.e., number and dose of immuno- suppressive agents used) and ensuing inability to control Epstein–Barr virus infection by the immune system (Opelz and Do¨ hler, 2003). Cases of immunosuppression-related lymphoma have specific features: they are generally B-cell non-Hodgkin lymphoma (NHL), may present as nodal or extra-nodal tumors and may occur in unusual locations. They are frequently polymorphic and may demonstrate clonal integration of Epstein–Barr virus genome in malignant cells (Liebowitz,   1998).   The   majority   of   immunosuppression-related lymphoma develops within the first 2 years after transplantation (Opelz and Do¨ hler, 2003).

Any association between AD and lymphoma irrespective of treatment is unclear. Some authors have found an increased risk of hematological malignancies in patients with eczema (Gibson et al., 1976; Bernard et al., 1984; Linet et al., 1986; Cartwright et al., 1988; Eriksson et al., 1995). On the other hand, at least two  studies  have  observed  that eczema decreases the risk of NHL (Bernstein and Ross, 1992; Fabbro-Peray et al., 2001). Recently So¨ derberg et al. (2004) found an association between ‘‘eczema’’ during childhood and NHL in adulthood yielding a relative  risk (RR) of 2.3; (95% confidence interval (CI) 1.0–5.3) (So¨ derberg et al., 2004). Zhang et al. (2004) found an increased odds ratio (OR) of T-cell NHL among women with eczema (OR 2.5; 95% CI 1.1–5.7) (Zhang et al., 2004).

Since December 2000 and December 2001, respectively, two topical calcineurin inhibitors (tacrolimus ointment and pimecrolimus cream) are available for the treatment of AD in the USA; the drugs are also available in other regions of the world. Long-term continuous application of tacrolimus ointment and pimecrolimus in an experimental ethanol solution to the skin of mice was associated with the development of lymphoma. This occurred at 26- and 47-fold higher exposures, measured by the area under the curve, respectively than the maximum individual exposure ever measured in humans after topical treatment (http://www. pharma.us.novartis.com/product/pi/pdf/elidel.pdf. Last vis- ited on 16 April 2006). The clinical relevance of the animal findings is uncertain because the skin of mice is more permeable than that of humans and high systemic exposure following topical application in rodents has been documen- ted whereas exposure in humans after topical application of calcineurin inhibitors is minimal (Harper et al., 2005; Paul et al., 2006). During post-marketing surveillance of topical tacrolimus and pimecrolimus, cases of lymphoma have been reported in patients treated with topical calcineurin inhibitors used alone or in combination with topical corticosteroids (Ormerod, 2005).We performed a case–control study to assess the risk of lymphoma associated with the use of topical prescription treatments for AD.

RESULTS

Description of the AD cohort
There were 293,253 patients of whom 171,724 (58.6%) were o20 years old (Table 1). Most patients (75.4%) were enrolled in the database from 2001 onwards. Almost 60% were female subjects. AD was mainly diagnosed by a family physician, pediatrician, or dermatologist. Twenty percent of patients met the definition of severe.

The OR (95% CI) for each of the covariates on the risk of lymphoma are shown on Table 2. Twenty-five percent of patients used topical steroids at AD index date compared to 3 and 1.5% using pimecrolimus and tacrolimus, respectively.

During follow-up, almost half of the patients used no medication, 40% used topical corticosteroids alone, and 12% were exposed to at least one of the topical calcineurin inhibitors. Of those exposed to the topical calcineurin inhibitors, 7.4% were exposed to pimecrolimus, 3.7% were exposed to tacrolimus, and 0.9% to both agents.

A total of 294 lymphomas occurred after the index date, 81 (27.6%) in patients o20 years old. The incidence rate of lymphoma was 81 per 100,000 person-years. The number of cases exposed to pimecrolimus, tacrolimus and both were 14, 11, and five, respectively. The type of lymphoma could not be specifically determined in 66% of cases. Among those that were identified Hodgkin disease accounted for 11.2%, and a NHL in 22.8%. B-cell NHL accounted for 4.4 and T-cell NHL for 18.4% of all lymphoma cases. All T-cell NHL except one were mycosis fungoides (MF) and the remaining case was Se´zary syndrome, which is commonly grouped with MF in the cutaneous T-cell NHL category.

OR for AD patients older than 30 years were elevated, especially for patients older than 60 years (OR 9.7; 95% CI 4.7–19.8). The risk of lymphoma was similar in male and in female subjects. Patients in the East region of the USA were found to have a higher risk of lymphoma (OR 2.1; 95% CI 1.5–2.9) compared to those in the Midwest. Severe AD was associated with an increased risk of lymphoma (OR 3.1; 95% CI 2.1–4.5). Visits to a dermatologist or an allergist in comparison to visits to a family physician at index date yielded elevated OR, as was the use of oral steroids during follow-up OR 1.9 (95% CI 1.4–2.7) or use of grade 1 (‘‘super potent’’) topical steroids at index date OR 2.2 (95% CI 1.3–4.0).

Table 3 presents the OR (95% CI) of lymphoma per treatment group with non-use as reference group. The results are presented unadjusted and adjusted for age, sex, region, medical specialty at AD diagnosis, presence of infectious mononucleosis, use of asthma medication, oral steroid use, and severity of AD. In the adjusted analysis, the OR for lymphoma associated with the different treatment modalities were: topical steroids high potency OR 1.2 (95% CI 0.8–1.8);  topical  steroids  low  potency  OR  1.1  (95%  CI  0.7–1.6); pimecrolimus OR 0.8 (95% CI 0.4–1.6), tacrolimus OR 0.8 (95% CI 0.4–1.7), and the concomitant use of topical steroids, pimecrolimus and tacrolimus OR 1 (95% CI 0.3–4.1) (the concomitant use of tacrolimus and pimecrolimus without steroids had only one case in the exposed group and the OR yielded an N result). Severe AD remained associated with an increased risk  of lymphoma (adjusted  OR  2.4;  95% CI 1.5–3.8), as did use of oral steroids (adjusted OR 1.5; 95% CI 1.0–2.4). High exposure to topical medications (topical steroids and/or topical calcineurin inhibitors remained associated with an increased risk of lymphoma (OR 2.30; 95% CI 1.17–4.51)).

In order to study the effect of duration of therapy on the relationship between treatment and lymphoma, and to avoid any risk of overmatching we repeated the analyses without matching cases and controls but adjusting for length of follow-up in a  logistic regression model.  We  repeated the stratified analysis using unconditional logistic regression and including length of follow-up as a confounding variable. The results were comparable to the matched results (data not shown). The risk of lymphoma increased with the length of follow-up as may be expected. Therefore we accepted the matched results.

DISCUSSION

The incidence rate of lymphoma was 81 per 100,000 person- years. This figure is compatible with an approximately 2.5- fold increase in risk in AD patients reported by other authors (So¨ derberg et al., 2004; Zhang et al., 2004). This study did not find an increased risk of lymphoma among users of topical calcineurin inhibitors. However, this finding should be confirmed when longer exposures to these drugs are available. The increase in risk associated with topical steroids disappeared in the adjusted analysis and the OR was slightly greater than one only in patients exposed to ‘‘high potency’’ topical steroids. This finding is difficult to interpret but, together with the higher OR observed for topical calcineurin inhibitors with concomitant use of topical steroids compared to topical calcineurin inhibitors alone, as well  as  the increased risk observed with oral steroids use suggests that further investigation of the relation between topical steroids and the occurrence of lymphoma in patients with AD may also be warranted. The use of systemic corticosteroids has previously been associated with an increased risk of lymphoma (Kato et al., 2002; Sorensen et al., 2004; Zhang et al., 2004), although other studies have not shown an increase (Beiderbeck et al., 2003; Holly and Bracci, 2003; Askling et al., 2005).

Severity of AD was associated with a 3-fold increase in the risk of lymphoma. Although other factors related to severity may play a role in this increased risk, the finding is consistent with the fact that other severe forms of chronic inflammatory diseases, such as rheumatoid arthritis and psoriasis, are associated with an increased risk of lymphoma (Margolis et al., 2001; Mariette et al., 2002; Gelfand et al., 2003). In the univariate analysis, high exposure category had an OR of 4.16 versus the non-exposed (95% CI 2.29–7.56), and severe AD had an OR of 3.08 (95% CI 2.11–4.49). It is virtually impossible to disentangle the effects of severity of disease on  outcome versus the true effects of drugs. However, in the adjusted analysis, that included severity and drug exposure, including topical steroids, in the same model, only severity of AD remained associated with a 3-fold increase in the risk of lymphoma.

It is possible that misclassification of cases or surveillance bias are partially responsible for the increased risk found in patients with severe AD. The high proportion of MF among cases makes misclassification a plausible hypothesis. MF and Se´zary syndrome can present with cutaneous manifestations that resemble chronic severe adult AD (Hagstromer et al., 2005) and it is possible that some of the cases diagnosed as AD with ensuing MF were in fact MF cases that had been misdiagnosed as AD. This is suggested by the high prevalence of MF among patients with very short (1–2 days) or short (o60 days) latency between index date and diagnosis of lymphoma (40–50% of all cases with very short latency were MF). It is also possible that severe cases were under the care of dermatologists and would be biopsied more frequently. The final analysis included any cases of lymphoma developing after  the  index  date.  Rather  than  applying  an  arbitrary threshold for making the latency time between exposure and lymphoma development plausible, we included all cases regarding their latency time. We repeated the analysis including only patients with at least  60  days  of  follow-up and the results were essentially unchanged (data not shown).

If topical immunomodulatory agents were to increase the risk of lymphoma it would be secondary to systemic immunosuppression. The most common presentation of lymphoma associated with systemic immunosuppression is a diffuse large cell B-cell NHL that is Epstein–Barr virus related and occurs relatively early in the first year (often the first 6 months) after solid organ transplantation (Swinnen, 2001; Faye and Vilmer, 2005; Ghobrial et al., 2005). Only one case, a Burkitt’s NHL, fitting this pathologic description is reported here and that case occurred 331 days after initiation of topical steroids alone. The remaining cases of B-cell NHL identified in this database are all described as chronic lymphocytic leukemia. Chronic lymphocytic leukemia is essentially unreported as a consequence of systemic im- munosuppressive therapy (de Lima et al., 1998; Diehl et al., 1999; Faye and Vilmer, 2005). In contrast among the cases with a definite histopathologic diagnosis, cutaneous T-cell NHL was the most frequent. Although skin T-cell lymphoma has been reported in association with significant immuno- suppression (Draoua et al., 2004), it is rare and tends to present late, often after more than a year of intense immunosuppression. The cases of MF observed in this study occurred remarkably early during follow-up.

The main limitation of the study was the inability to validate information obtained by record linkage in Phar- Metrics. Therefore, we were unable to ascertain the degree of misclassification that may have occurred. The selection of controls was random and the hematologist who classified the cases was blinded to treatment group, so misclassification should be randomly distributed among the different treatment groups. In addition, the misclassification was suspected to mainly consist of false positive cases of MF (cases of severe AD whose eczematoid cutaneous expression would be misclassified as MF). Therefore, reducing misclassification might result in a reduced OR for topical steroids but it would be unlikely to yield higher OR with other treatment groups. We repeated the analysis excluding the MF/T-cell NHL cases and the results did not change (Table 3). The direction of the minimal changes observed was either towards the null value of the OR or a reduced risk associated with the use of topical agents. Likewise, lack of validation did not allow studying the relationship between different subtypes of lymphoma and topical AD treatments.

Another limitation of the study is that information on over the counter (OTC) drug use, including OTC topical steroids, is lacking in PharMetrics. OTC topical steroids are of low potency with no anticipated effect on lymphoma risk, according to the results of this study. The prevalence of severe AD was higher in patients receiving calcineurin inhibitors and there was no increased risk of lymphoma among these patients (with and without steroids). Therefore it seems that the potential bias introduced by OTC topical steroid use would not modify the results substantially, if at all.
 
The effect of missing OTC drug use in epidemiologic research has been considered negligible in other studies (Drews and Greenland, 1990).

Finally, this study was based on relatively small number of cases and with relatively short follow-up and exposure times. This was, to some extent, unavoidable given that lymphoma is a rare disease, especially among younger patients, AD treatment is typically intermittent and topical calcineurin agents have been recently introduced. The case–control design used can be conceptualized as a follow-up design in which the person time experience of the denominators of the incidence rate is sampled rather than measured outright. It is also important to highlight that the selection of cases and controls was ‘‘nested’’ in a cohort of patients with AD in order to decrease any possible effect that this condition may have in the risk of lymphoma.

The large sample of AD patients used is the main strength of this study. By performing the analysis nested in a cohort of patients with AD we decreased the possibility of confounding by indication, detecting an increased risk of lymphoma owing to AD itself irrespective of treatment. It is possible that some residual confounding remains (e.g., by severity of AD). As some of the AD patients were not being treated at the index date it could be argued that using a more stringent definition of AD might have yielded different results. However, we  adjusted  for  severity in the  final  model  and also repeated the analyses restricting inclusion to those patients with prescription  topical treatment  of AD and  the results did not change (data not shown). The absence of treatment at index date is not surprising. Patients with AD, even severe, are not treated continuously for a variety of reasons, including patients’ or parents’ fear of steroids (e.g., keeping fluorinated drugs off facial skin of a child), cost of medications, desire to try emollients first, or concerns about compliance.

This study showed no increased risk of lymphoma in patients treated with topical calcineurin inhibitors. It is difficult to disentangle the effects of severity of disease on outcome versus the true effects of drugs. However, in the adjusted analysis, severity of AD was the main factor associated with an increased risk of lymphoma. It would be important to confirm this finding allowing longer exposure to these drugs in a population that allows medical validation of cases. Likewise it would be important to evaluate the role of severity of AD, high potency topical steroids, and oral corticosteroids in the development of lymphoma in patients with  AD.

MATERIALS AND METHODS

The study population was derived from the PharMetrics database, which includes data from 43 million US patients from 73 health-care plans (http://www.pharmetrics.com/p_overview.html Last visited March 5, 2005). Records in the PharMetrics database are represen- tative of the USA managed care population, based on a variety of patient and health plan demographic measures that include geographic region, age, gender, and plan type.

We obtained all data on patients with International classification of diseases codes 691 and 692 from July 1995 to January 2005. We  then extracted information on all subjects with a history of AD, indicated by the presence of an International classification of diseases-9 CM code consistent with AD: 691.8 (dermatitis other atopic) and 691 (dermatitis atopic) but not 691.0 (rash diaper or napkin). We considered the index date for each patient the day a code for AD was first present in the database.

The population included in PharMetrics consisted  of  502,283 with AD. We limited inclusion in the cohort to patients with at least 6 months of enrollment in the database. We then excluded patients with the diagnoses of lymphoma, cancer, immunosuppression, transplantation, HIV infection and/or AIDS, use of immunosuppres- sive agents, and anticancer drugs before the index date. After applying these inclusion and exclusion criteria the final cohort included 293,253 patients with AD.

Description of the AD cohort
Patients in the cohort were characterized according to the following parameters: age, gender, specialty of the treating physician at index date, year of entry in the cohort based on AD index date, geographic region (East, Midwest, South, West), presence of mononucleosis  infection,  or  asthma  during   the   follow-up period, use of asthma medication and oral corticosteroids at the index date, topical corticosteroid use at the index date, pimecroli- mus and tacrolimus use at the index date, and severity of AD. Severity of AD  was  defined  based  on  criteria  from  Margolis et al. (2001) as: patients who experienced at least four physician visits for AD per year (for patients 3 years and older) or if they had had at least one visit to a dermatologist in  the  follow-up period (for patients younger than 3 years). Topical corticosteroid potency was classified from grade 1 ‘‘superpotent’’ to grade  7 ‘‘least potent’’ according to  a  modified  version  of  the  USA National Psoriasis Foundation classification of steroid potency. (Potencies of topical steroids. National Psoriasis Foundation. http://www.psoriasis.org/treatment/psoriasis/steroids/potency.php? PHPSESSID=5cb7cdba9997d740f09bb6534f791e38. Last visited on 10 January 2006.)

Cases and controls identification and characterization
Cases were identified by the presence of a code for lymphoma (International classification of diseases codes 200, 201, 202, and 204) after the AD index date. The cases of lymphoma were reviewed by a hematologist blinded to exposure. Cases were subdivided based on diagnostic coding as: Hodgkin disease, NHL, and indeterminate (meaning there was insufficient information to make a finer distinction). NHL included B-cell NHL (specific diagnoses provided were Burkitt’s lymphoma and chronic lymphocytic leukemia) and T-cell NHL (specific diagnoses provided were MF and Se´zary syndrome). Where there were simultaneous diagnoses of indetermi- nate lymphoma and a more specific descriptor (e.g., Hodgkin disease or MF) on the same day, only the more specific diagnosis was recorded for classification. In one case, two diagnoses were recorded for the same patient several years apart and two different diagnoses were used, acknowledging the finding could be a second malignancy.

Controls were patients from the AD cohort who did not have a code for lymphoma at the time lymphoma was diagnosed in a matched case. Cases and controls were matched for duration of follow-up calculated from the index date until a lymphoma was  diagnosed in the cases according to the risk set sampling method (Navidi and Weinhandl, 2002). For each case, four controls having at least the same duration of follow-up were randomly selected from the cohort of AD  patients. Patients who subsequently  developed lymphoma were eligible to serve as controls until the date of their lymphoma diagnosis. An individual could serve as a control for more than one case.

Categorization of medication exposure
Exposure to topical AD immunomodulatory treatment was categor- ized based both on intensity of exposure and type of medication using National Drug Codes.

Intensity of exposure. Patients in the cohort were classified as belonging to ‘‘no-use’’, ‘‘low’’, or ‘‘high’’ exposure categories for topical AD prescription drugs: topical calcineurin inhibitors and topical corticosteroids. Exposure time was estimated from the number of days for which treatment was supplied on relevant pharmacy claims. Patients were considered to be ‘‘persistent users’’ of pimecrolimus, and/or tacrolimus, and/or topical corticosteroids as long as the window between two courses of therapy did not exceed twice the number of days supplied on the previous claim. Upon not fulfilling this persistence definition, the patient was considered discontinued from their index therapy as of the date when the most recent prescription in the course of therapy was ‘‘exhausted’’, this date was deducted from the number of days for which the medication was supplied.

Patients o3 years old were classified as high exposure users if at least a third of their follow-up time was covered by pimecrolimus, tacrolimus, or  topical corticosteroids.  Patients X3 years old  were classified as high exposure users if they had at least three prescriptions of pimecrolimus or tacrolimus per year or they had at least three prescriptions of topical corticosteroids classified as potency p3 per year.

All patients having at least 1 day of their follow-up time covered by calcineurin inhibitors or topical corticosteroid and who did not meet the criteria for high exposure were classified as low exposure users.

Patients who did not use a calcineurin inhibitor or a topical corticosteroid during their follow-up time were considered ‘‘no-use’’ exposure patients.

Type of medication used: exposure group allocation. Cases and controls  were allocated to one of five  mutually  exclusive exposure groups depending on the type of drug used during the follow-up period: (1) topical steroids if there was a prescription for a topical steroid and not for a calcineurin inhibitor, (2) pimecrolimus if there was a prescription for pimecrolimus regardless of the presence of a prescription for topical steroid, (3) tacrolimus if there was a prescription for tacrolimus regardless of the presence of a prescrip- tion for topical corticosteroid, (4) both (tacrolimus and pimecroli- mus) if there was a prescription for pimecrolimus and tacrolimus regardless of the presence of a prescription for topical corticosteroid,

(5) none, if there was no prescription for a topical corticosteroid or for a topical calcineurin inhibitor. Patients classified under the ‘‘pimecrolimus’’, ‘‘tacrolimus’’, and ‘‘both’’ categories were split into those who received concomitant topical steroids and those who did not.
 
Analysis
The relationship between exposure to medications and lymphoma occurrence was first described in cross-tabulations. We used logistic regression conditional on case sets with similar duration of follow-up to calculate OR and 95% CI. We  described the  covariates distribution among controls classified according to the different levels of exposure to medications to analyze differences among treatment groups. The final model was adjusted for all confounders with an effect 410% on the estimate and confidence limits accrued with all  covariates in the initial model (Greenland and  Rothman, 1998).

CONFLICT OF INTEREST

This study was sponsored by Novartis, the manufacturer of one of the calcineurin antagonists studied. Risk Management Resources had a contrac- tual agreement giving Risk Management Resources complete control of the content of this paper as well as to the journal of choice for submission. Felix M. Arellano has been a paid consultant for Novartis. Carlos Ferna´ndez- Vidaurre is a Novartis employee. Carle Paul was an employee of Novartis at the time the study was performed.

ACKNOWLEDGMENTS

We are indebted to Dr Eva Guinan, Dana-Farber Cancer Institute, Boston, for review of the cases of lymphoma and to Barbara Rosin, Novartis Pharma for editorial assistance.
 

REFERENCES

1. Askling J, Klareskog L, Hjalgrom H, Baecklund E, Bjo¨ rkholm M, Ekbom A (2005) Do steroids increase lymphoma risk? A case–control study of lymphoma risk in polymyalgia rheumatica/giant  cell  arteritis.  Ann Rheum Dis 64:1765–8
2. Beiderbeck AB, Holly EA, Sturkenboom MC, Coebergh JW, Stricker BH, Leufkens HG (2003) No increased risk of non-Hodgkin’s lymphoma with steroids, estrogens and psychotropics (Netherlands). Cancer Causes Control  14:639–44
3. Bernard SM, Cartwright RA, Bird CC, Richards ID, Lauder I, Roberts BE (1984) Aetiologic factors in lymphoid malignancies: a case–control epidemio- logical study. Leuk Res 8:681–9
4. Bernstein L, Ross RK (1992) Prior medication use and health history as risk factors for non-Hodgkin’s lymphoma: preliminary results from a case–- control study in Los Angeles County. Cancer Res 52(Suppl 19): 5510s–515s
5. Cartwright RA, Darwin C, McKinney PA, Roberts B, Richards ID, Bird CC (1988) Acute myeloid leukemia in adults: a case–control study in Yorkshire. Leukemia 2:687–90
6. de Lima M, O’Brien S, Lerner S, Keating MJ (1998) Chronic lymphocytic leukemia in the young patient. Semin Oncol 25:107–16
7. Diehl LF, Karnell LH, Menck  HR  (1999)  The  American  college  of surgeons commission on cancer and the American cancer society. The national cancer data base report on age, gender, treatment, and outcomes of patients with chronic lymphocytic leukemia. Cancer 86:2684–92
8. Draoua HY, Tsao L, Mancini DM, Addonizio LJ, Bhagat G, Alobeid B (2004) T-cell post-transplantation lymphoproliferative disorders after cardiac transplantation: a single institutional experience. Br J Haematol 127: 429–32
9. Drews CD, Greenland S (1990) The impact of differential recall on the results of case–control studies. Int J Epidemiol 19:1107–12
10. Eriksson NE, Holmen A, Hogstedt B, Mikoczy Z, Hagmar L (1995) A prospective study of cancer incidence in a cohort examined for allergy. Allergy  50:718–22
11. Fabbro-Peray P, Daures JP, Rossi JF (2001) Environmental risk factors for non- Hodgkin’s lymphoma: a population-based case–control study in Langue- doc-Roussillon, France. Cancer Causes Control 12:201–12
12. Faye A, Vilmer E (2005) Post-transplant lymphoproliferative disorder in children: incidence, prognosis, and treatment options. Paediatr Drugs 7:55–65
13. Fleming S, Bodner C, Devereaux G, Russell G, Campbell D, Godden D et al. (2001) An application of the United Kingdom working party diagnostic criteria for atopic dermatitis in Scottish infants. J Invest Dermatol 117:1526–30
14. Gelfand JM, Berlin J, Van Voorhees A, Margolis DJ (2003) Lymphoma rates are low but increased in patients with psoriasis: results from a population-based cohort study in the United Kingdom. Arch Dermatol 139:1425–9
15. Ghobrial IM, Habermann TM, Macon WR, Ristow KM, Larson TS, Walker RC et al. (2005) Differences between early and late post transplant lymphoproliferative disorders in solid organ transplant patients: are they two different diseases? Transplantation 79:244–7
16. Gibson R, Graham S, Lielienfeld A, Schuman L, Levin M, Swanson M (1976) Epidemiology of diseases in adult males with leukemia. J Natl Cancer Inst 56:891–8
17. Greenland S, Rothman KJ (1998) Introduction to stratified analysis. In: Modern Epidemiology (Rothman KJ, Greenland S eds), 2nd ed. Philadelphia: Lippincot, Raven, 253–79
18. Hagstromer  L,  Ye  W,  Nyren  O,  Emstestam  L  (2005)  Incidence  of cancer among patients with atopic dermatitis. Arch Dermatol 141: 1123–7
19. Harper J, Smith C, Rubins A, Green A, Jackson K, Zigure S et al. (2005) Multicenter study of the pharmacokinetics of tacrolimus ointment after first and repeated application to children with atopic dermatitis. J Invest Dermatol  124:695–9
20. Holly   EA,   Bracci   PM   (2003)   Population-based   study   of   non-Hodgkin lymphoma, histology, and medical history among human immunodefi- ciency virus-negative participants in San Francisco. Am J Epidemiol 158:316–27
21. Kato I, Koenig KL, Shore RE, Baptiste MS, Lillquist PP, Frizzera G et al. (2002) Use of anti-inflammatory and non-narcotic analgesic drugs and risk of non-Hodgkin’s lymphoma (NHL) (United States). Cancer Causes Control 13:965–74
22. Liebowitz D (1998) Epstein–Barr virus and a cellular signaling pathway in lymphomas from immunosuppressed patients. N Engl J Med 338: 1413–21
23. Linet MS, McCaffrey LD, Humphrey RL, Brookmeyer R, Van Natta ML, Tielsch JM et al. (1986) Chronic lymphocytic leukemia and acquired disorders affecting the immune system: a case–control study. J Natl Cancer Inst 77:371–8
24. Margolis D, Bilker W, Hennessy S, Vittorio C, Santanna J, Strom BL (2001) The risk of malignancy associated with psoriasis. Arch Dermatol 137: 778–83
25. Mariette X, Cazals-Hatem D, Warszawki J, Warszawki J, Liote F, Balandraud N, Sibilia J et al. (2002) Lymphomas in rheumatoid arthritis patients treated with methotrexate: a 3-year prospective study in France. Blood 99:3909–15
26. Navidi W, Weinhandl E (2002) Risk set sampling for case-crossover designs.
27. Epidemiology  13:100–5
28. Opelz G, Do¨ hler B (2003) Lymphomas after solid organ transplantation: a collaborative transplant study report. Am J Transpl 4:222–30
29. Opelz G, Henderson R (1993) Incidence of non-Hodgkin lymphoma in kidney and heart transplant recipients. Lancet 342:1514–6
30. Ormerod AD (2005) Topical tacrolimus and  pimecrolimus  and  the risk of cancer: how much cause for concern? Br J Dermatol 153: 701–5
31. Paul C, Cork MJ, Rossi AB, Papp K, Barbier N, de Prost Y (2006) Safety and tolerability of pimecrolimus cream 1% in infants: experience in 1133 patients treated for up to 2 years. Pediatrics 117:e118–28
32. So¨ derberg KC, Hagmar L, Schwartzbausm J, Feychting M (2004) Allergic conditions and risk of hematological malignancies in adults: a cohort study. BMC Public Health 4:51 doi:10.1186/1471-2458-4-51
33. Sorensen HT, Mellemkjaer L, Nielsen GL, Baron JA, Olsen JH, Karagas M (2004)  Skin  cancers  and  non-Hodgkin  lymphoma  among  users  of  systemic  glucocorticoids.  A  population-based  cohort  study.  J   Natl Cancer Inst 96:709–11
34. Su JC, Kemp AS, Varigos GA, Nolan TM (1997) Atopic eczema: its impact on the family and financial cost. Arch Dis Child 76:159–62
35. Swinnen LJ (2001) Post-transplant lymphoproliferative disorders: implications for   acquired   immunodeficiency   syndrome-associated   malignancies. J Natl Cancer Inst Monogr 28:38–43
36. Williams H, Robertson C, Stewart A, Ait-Khaled N, Anabwani G, Anderson R
37. et al. (1999) Worldwide variations in the prevalence of symptoms of  atopic  eczema  in  the  international  study  of  asthma  and  allergies  in childhood. J  Allergy Clin Immunol 103:125–38
38. Williams HC (1997) Inflammatory skin diseases 1: atopic dermatitis. In: The Challenge of Dermato-Epidemiology (Williams HC, Strachan DP, eds), Boca Raton: CRC Press Inc, 125–44
39. Zhang Y, Holford TR, Leaderer B, Zahm SH, Boyle P, Morton LM et al. (2004) Prior medical conditions and medication use and risk of non-Hodgkin lymphoma in connecticut United States women. Cancer Causes Control 15:419–28

Вопросы, ответы, комментарии