Recognizing Primary Immune Deficiency in Clinical Practice

Primary immunodeficiency results in recurrent infections, organdysfunction, and autoimmunity. We studied 237 patients referredfor suspicion of immunodeficiency, using a scoring system basedon clinical information. The 113 patients with immunodeficiencyhad higher scores and more episodes of chronic illnesses andwere more likely to have neutropenia, lymphopenia, or splenomegaly.

INTRODUCTION

Top

Introduction

Primary immunodeficiency (PI) diseases may be recognized dueto sinopulmonary or gastrointestinal tract infections, organabscesses, autoimmunity, or systemic signs, such as fever orfailure to thrive (3, 6, 11, 14, 32). However, because of thediversity of immune defects, range of ages, and different clinicalmanifestations, distinguishing patients with PI in practicecan be challenging. While delays in diagnosis are common (12,19, 32), studies have shown an 8 to 24% incidence of immunodeficiencyin selected patient populations (7, 8, 12, 19, 24, 25, 29).

We previously used a computerized method to identify potentiallyimmunodeficient patients repeatedly hospitalized due to illnessescharacteristic of PI (15); 30% of a cohort of these subjectshad immune defects (15). In the present study, we evaluateda group of physician-referred patients because of conditionssuggestive of immunodeficiency. Our goal was to determine ifpatients shown to have immune defects had differences from patientswho did not.

MATERIALS AND METHODS

Top

Materials and Methods

Study participants and testing. Adult or pediatric patients who were referred by a physicianbetween 1 January 2001 and 1 July 2003 were asked to participatein the study. Laboratory studies included analyses of immuneglobulins and antibody responses to vaccines (tetanus, diphtheria,pneumococcus [12 serotypes], Haemophilus influenzae) and isohemagglutinintiters (13, 23). Diagnosis of immunoglobulin G1 (IgG1) or IgG2subclass deficiency was defined as IgG1 or IgG2 levels 2 standarddeviations less than the age-related means for individuals withantibody deficiencies (2). Lymphocyte surface markers, lymphocyteproliferation, neutrophil reduction of dihydrorhodamine, andcomplement testing were done by standard assays (13, 16, 23,30). A diagnosis of immune deficiency was made by the use ofstandard criteria (10).

Historical information. The clinical histories for a 5-year period were reviewed. Diagnoseswere confirmed by cultures, radiographic procedures, or thereview of the records from the referring physicians. As describedpreviously, illnesses suggestive of PI (4, 9, 12, 18, 20, 22,27, 28, 31, 32) (Table 1) were recorded on a case report form.These illnesses or conditions were arbitrarily assigned a scoreof "1", "2," or "3" to indicate relative severity (15). Chronicconditions were counted only once in a year's time. An immunodeficiency-related(IDR) score was the sum of the scores acquired over the previous5 years. To account for the various ages of the patients, thenumber of years of observation was expressed as the number of"person years" of observation.

View this table:
[in this window]
[in a new window]
TABLE 1. Scores of diagnoses and conditions assessed from the clinical record

Statistical analysis. Chi-square analysis was used to compare the IDR scores for bothgroups. The numbers of conditions in both groups were comparedby using age-adjusted time ("person years") and the Mann-WhitneyU test (SPSS statistical software; SPSS, Inc., Cary, NC). Amultivariate analysis was performed by evaluating each diagnosisand its association with PI. A P value of less than 0.05 wasconsidered significant.

RESULTS

Top

Results

From 1 January 2001 to 1 July 2003, 237 patients were referredby internists, pediatricians, or allergists for suspected immunodeficiency.Of the patients referred, 52% were females and 72% were Caucasians.The median age was 24.5 years (range, 1 to 85 years); immunodeficiencywas diagnosed in 113 patients (48%) (Table 2).

View this table:
[in this window]
[in a new window]
TABLE 2. Characteristics of patient groups^a

For patients with immunodeficiency, the most common diagnosiswas common variable immunodeficiency (32%), followed by IgGsubclass deficiency and IgA deficiency (Table 3). For the 124patients not found to have an immune defect, the reasons forreferral included chronic bronchitis, chronic sinusitis, pneumonia,and acute or chronic otitis media. These patients had asthma(n = 11), allergic rhinitis and/or atopic dermatitis (n = 8),or food allergies (n = 2). Twenty-three percent of the patientshad autoimmune or inflammatory diseases, including rheumatoidarthritis (n = 4), unspecified autoimmunity (n = 3), systemiclupus erythematosus (n = 1), or fibromyalgia rheumatica (1).Other subjects had Crohn's disease (n = 3); diabetes mellitus(n = 2); chronic warts (n = 2); or (one each) quadraplegia,swallowing disorder, cleft palate, osteogenesis imperfecta,Gaucher's disease, sarcoidosis, Epstein-Barr virus-related lymphohistiocytosis,obsessive compulsive disorder, alopecia areata, Hirshprung'sdisease, familial Mediterranean fever, growth hormone deficiency,or ulcerative colitis.

View this table:
[in this window]
[in a new window]
TABLE 3. Diagnoses for patients with identified immune deficiency^a

Immune deficiency-related scores and specific medical conditions. Patients with immunodeficiency had a median score of 8 (interquartilerange, 5 to 13), which was significantly higher (P = 0.004)than the median score for those who did not (median score, 6;interquartile range, 3 to 10). By the use of age-adjusted comparisons,patients with immune deficiency had a median score of 10 (interquartilerange, 5 to 17.5), which was higher (P = 0.025) than the medianscore of those without immune deficiency (median score, 8; interquartilerange, 4 to 14). However, an IDR score of 8 or greater had apositive predictive value of 59%. By consideration of the patients'medical conditions, patients with immunodeficiency had significantlymore person years of chronic sinusitis, chronic bronchitis,chronic otitis media, and chronic diarrhea than patients withoutimmune defects (P = 0.001, 0.001, 0.001, and 0.001, respectively)(Table 4). Lymphopenia, thrombocytopenia or neutropenia, andsplenomegaly were also more characteristic of immunodeficiency(P = 0.011, 0.027, and 0.011, respectively). In multivariateanalyses, splenomegaly was associated with immunodeficiency(odds ratio [OR] = 7.9), followed by neutropenia (OR = 5.0),chronic diarrhea (OR = 2.5), and chronic sinusitis (OR = 1.8).Acute otitis media and suppurative otitis were found more frequentlyin patients without immune defects (P = 0.001 and 0.012, respectively).

View this table:
[in this window]
[in a new window]
TABLE 4. Comparison of diagnoses and conditions for patient groups

DISCUSSION

Top

Discussion

Primary immunodeficiency diseases are relatively rare disorders,but patients with such diseases may appear in general clinicalpractice. While severe immune defects are more easily recognized,milder defects may not be diagnosed until illness or hospitalizationoccurs. However, determination of which patients should be evaluatedis not always clear, since conditions that occur in immunodeficientsubjects are common in subjects with healthy immune systems.

We previously used a scoring system based on the codes of theInternational Classification of Diseases, version 9 (34), forhospitalized subjects (15) to identify potentially immunodeficientpatients. A minimum score of 6 allowed us to identify subjectswith immune defects, although most subjects had higher scores(15). In the present study we used a similar method, which weapplied to the medical records of patients referred by otherphysicians. While patients with immunodeficiency had a highermedian score than those without immunodeficiency, due to theclinical overlap of conditions, the sensitivity of the IDR scoresalone to the identification of PI was low. However, chronicsinusitis, bronchitis, otitis media, and chronic diarrhea werefound more often in the immunodeficient group. Note that whilepneumonia is characteristic of immunodeficiency (1, 5, 12, 21,32, 33), this condition was common in both groups.

In a prior report, only 10% of 238 children with recurrent pneumoniahad immune defects (24), suggesting that while pneumonia isa useful hallmark of PI, it is a nonspecific indicator.

A scoring system previously proved useful for the identificationof immunodeficient subjects in a cohort of patients who hadhad multiple hospitalizations; a similar scoring system coupledwith specific clinical indicators may provide a useful guideto the identification of such patients in the outpatient setting.

ACKNOWLEDGMENTS

This work was supported by grants AI-101093, AI-467320, andAI-48693 from the National Institutes of Health and NIAID contract03-22.

FOOTNOTES

* Corresponding author. Mailing address: Division of Clinical Immunology, Mount Sinai Medical Center, 1425 Madison Ave., New York, NY 10029. Phone: (212) 659-9261. Fax: (212) 987-5593. E-mail: hale.yar{at}mssm.edu.

REFERENCES

Top