Introduction
In the Fall of 1997, increased demand, reduced supply, and product recalls created a shortage of intravenous immunoglobulin (IVIG) in the United States. Key factors in the shortfall in production included issues of regulatory compliance and good manufacturing practices. In response, the Food and Drug Administration (FDA) facilitated production and distribution, shortened the review of lot release protocols, expedited the review of license applications, and streamlined clinical trial design.
At the FDA Blood Products Advisory Committee (BPAC) meeting held in March, 1999 [
1], the FDA recommended a change in design for a pivotal clinical trial to evaluate the safety and efficacy of a new IVIG product. The new design involved a prospective, randomized, double-blind, parallel-group, positive-controlled, non-inferiority study in 80 subjects with a documented history of primary immune deficiency, in which the safety and efficacy of the test product was to be compared head-to-head to a US-licensed IVIG product. Investigational new drug (IND) sponsors would evaluate efficacy by comparing the serious infection rate in each randomization group over an observation period of 12 months.
Following the March 1999 BPAC meeting, the FDA increasingly recognized several problems with this trial design, e.g., limited numbers of patients that could be recruited for trials, multiple new IVIG products to be tested, and the potential for IVIG shortages. At the March 2000 BPAC meeting [
2], the FDA presented an alternate clinical trial design for the evaluation of IVIG safety and efficacy in primary immune deficiency (PID). The new proposal was a single-arm, 12-month, open study of approximately 50 PID patients with safety and efficacy targets based on previous trials. Pharmacokinetic studies would be performed on at least 20 PID patients and observed values should not be inferior to those previously determined for approved products.
Since 1999, the FDA has issued two guidance documents for clinical trials of IVIG to support the marketing of IVIG as replacement therapy for primary humoral deficiency [
3,
4]. The 2005 FDA Guidance [
3] defined the primary efficacy endpoint as the rate of acute serious bacterial infections (SBIs) during the year of treatment with the study IVIG. Based on the FDA’s examination of historical data, the SBI rate was to be <1.0 per subject per year at the 0.01 level of significance. SBIs were defined as bacteremia/sepsis, bacterial meningitis, osteomyelitis/septic arthritis, bacterial pneumonia, and visceral abscess. Essential diagnostic features for each infection were described.
Safety endpoints were the overall incidence of adverse events (AEs) (regardless of causality) that occurred during or within 1, 24, and 48 h after infusion of the test product. Temporally associated AEs are those that occurred during infusion and up to 48 h after infusion. A primary endpoint was to record the proportion of infusions with one or more temporally associated AEs (including AEs that were determined to be unrelated to the product). The target for this endpoint was an upper one-sided 95 % confidence limit of <0.40 [
3]. The number of subjects to achieve the suggested endpoints was estimated at 40–50 subjects.
The 2008 FDA Guidance [
4] extended the time period for temporally associated AEs to 72 h and added instructions for conducting pharmacokinetic studies in pediatric patients.
The effect of the 2005 and 2008 Guidances has been to standardize endpoints for safety, efficacy, and pharmacokinetics in clinical trials conducted in the United States over the last decade. We report here the results of clinical trials in patients with primary immunodeficiency treated with IVIG products currently licensed in the United States.
Early studies
In 1890, Emil Behring and Shibasaburo Kitasato discovered that the injection of guinea pigs with sterilized cultures of diphtheria or tetanus bacilli produced substances in the sera that could neutralize diphtheria or tetanus toxins. They further showed that sera produced in one animal could be used to cure infected animals and to protect healthy animals from diphtheria or tetanus infection mortality [
5,
6].
In the 1930s, Enders showed that normal human serum could prevent measles [
7]. Contemporary attempts to concentrate this activity in the gamma globulin fraction often resulted in bacterial contamination and adverse reactions in recipients [
7]. The use of cold ethanol fractionation as a means to isolate gamma globulin from human plasma finally provided a consistently safe material for injection [
7‐
9]. This material, Cohn Fraction II, was shown by Enders to contain the majority of anti-measles activity [
7].
In 1952, Bruton treated a child that had undetectable gamma globulin levels and recurring pneumococcal infections with subcutaneous infusions of human immune serum globulin (ISG, i.e., Cohn Fraction II formulated at 165 mg/mL). The injection of 3.2 g/month ISG increased circulating gamma globulin levels and completely eliminated pneumococcal infections [
10]. These observations were rapidly confirmed [
7] and treatment with human IgG soon became the standard of care for patients with primary antibody deficiencies [
7].
Early clinical studies of ISG were conducted primarily in children who suffered from recurrent bacterial infections, exhibited evidence of deficient antibody production, and had extremely low levels of circulating gamma globulin [
11]. ISG was restricted to intramuscular injection because intravenous injection caused severe pyrogenic and cardiovascular reactions in many recipients [
12]. In response, a series of manufacturing changes designed to reduce the incidence of side effects led to the development of immune globulins intended for intravenous injection (IVIG).
Spontaneous complement activation (anti-complement activity) by IgG aggregates was considered to be the principal cause of adverse reactions [
13]. To suppress anti-complement activity, some manufacturers treated Cohn Fraction II with enzymes or chemical modification. Unfortunately, these treatments also reduced important antibody biological activities required for clinical efficacy and shortened IgG circulating half-lives [
14‐
16].
The recognition that Cohn Fraction II contained trace amounts of highly active contaminants such as prekallikrein activator, prekallikrein, and activated coagulation factors led to the development of Cohn Fraction II purification procedures using anion (DEAE) exchange chromatography. The first DEAE-purified IVIG contained none of the trace contaminants associated with AEs and some antibody biological activities, such as bacterial opsonization and virus neutralization, were higher than in “treated” products [
17]. Most commercial IVIGs are now produced with an anion exchange chromatography step and contain relatively low levels of trace contaminants.
In the 1980s, the licensure of IVIG in the United States was generally based on data from relatively small numbers of PID patients. The first IVIG licensed in the United States was chemically modified. Twenty patients with X-linked agammaglobulinemia (XLA,
n = 9) or common variable immunodeficiency (CVID,
n = 11) were randomized to receive either 150 mg/kg/4 weeks of IVIG or 100 mg/kg/4 weeks of ISG. Fourteen patients were treated with IVIG for 242 treatment months. This group developed 0.103 acute infections/month of treatment. In contrast, 13 patients treated with ISG for 193 treatment months developed 0.295 acute infections/month of treatment [
18].
Cunningham-Rundles et al. [
19] studied 21 patients who were treated with 300 mg/kg IVIG every 3 weeks for 1 year following a year of treatment with ISG. In the 18 patients who completed a year of IVIG treatment, the average IgG trough level was 505 mg/dL, compared to 262 mg/dL during ISG treatment. Efficacy endpoints such as days of illness from infection (e.g., upper respiratory tract infections, bronchitis, sinusitis, etc.) were significantly lower (
p < 0.001) during the IVIG treatment year compared to the ISG treatment year. Adverse reactions to IVIG were observed in 10 of 407 infusions (2.5 %) and were eliminated by slowing or temporarily stopping the infusion [
19].
In a trial of another IVIG, Ochs et al. treated 15 PID patients with 400 mg/kg every 4 weeks for 1 year. Twenty-one adverse reactions were associated with 15 of 341 infusions (4.4 %) during the first 48 h after infusion. The incidence of acute upper respiratory infections was as low or lower than infections reported for other preparations [
20].
In a third trial, Roifman and Gelfand compared high- versus low-dose therapy in patients with hypogammaglobulinemia and sinopulmonary disease. High-dose therapy (600 mg/kg every 4 weeks) proved efficacious in reducing symptoms, decreasing the frequency of major and minor infections, and significantly improving pulmonary function. The improvement appeared to correlate with a marked reduction in the isolation of
Mycoplasma, particularly
Ureaplasma urealyticum, an important cause of infection in patients with hypogammaglobulinemia [
21].
Discussion
The 1999 BPAC meeting concluded that each IVIG is unique and should not be treated as a single generic product [
1]. Consequently, the FDA proposed that phase III clinical trials of new IVIGs (including licensed products with modified manufacturing procedures) should be more rigorous than previous trials of small numbers of patients [
18‐
20]. It soon became apparent that the number of patients with primary immunodeficiency that could be recruited for IVIG clinical trials was too limited to support the number of new IVIGs that needed to be tested. This prompted the analysis of possible IVIG trials that would reduce the sample size.
A new proposal was to conduct a single-arm, 12-month, open-label study with comparison to historical controls for safety, pharmacokinetics, and efficacy endpoints. The FDA defined infusional AEs as those that are temporally associated with an IVIG infusion, i.e., occurred during or within 72 h of an infusion. The safety endpoint was defined as the proportion of infusions with one or more temporally associated AEs (including AEs that are determined not to be product-related). The target for this endpoint is an upper one-sided 95 % confidence limit of <0.40 [
4].
The efficacy endpoint is the incidence of SBIs. Based on historical data, patients with primary humoral immunodeficiency experienced approximately 4 or more SBIs per year prior to routine immunoglobulin therapy [
4]. Also, the SBI rate was <0.5 per year during periods of regular immunoglobulin therapy (200–600 mg/kg/infusion, every 3 or 4 weeks) [
4]. Accordingly, the FDA set an SBI target of an upper one-sided 99 % confidence limit <1.0 SBI per person per year. The criteria for diagnosing SBIs were defined in the Guidance [
4].
The safety and efficacy endpoints in the FDA IVIG Guidances [
3,
4] have essentially been incorporated into each of the IVIG clinical trials published since 2004. The trial that compared Gamunex 10 % to Gamimune N 10 % used serious infection endpoints that were different from the other studies.
Although many of the same study sites were involved in studies of multiple products, the patient population in each clinical trial is assumed to be different. However, the majority (52 to 92 %) of patients in each study were diagnosed with CVID. The proportion of XLA patients varied from 0 to 28 % (Table
1).
The rate of SBIs (the primary efficacy endpoint) in each study was well below the 0.5 per year reported by the FDA as the historical frequency during periods of regular immunoglobulin therapy [
4]. SBI rates varied from 0 to 0.08 SBI/patient/year (Table
1). The rate of non-serious infections also varied from study to study, with values of 1.96–3.55 infections/patient/year (Table
1).
Days of work and school missed because of PID-related illness were reported in six studies and varied from an average of 2.1 days/year (Biotest-IVIG) to 13 days/year (Flebogamma 5 %). The values of 13 and 8.7 days of missed work/school resulted from relatively few patients. The average number of hospitalization days per year caused by PID-related infections was low in each study, ranging from 0.21 (Biotest-IVIG) to 2.31 (Privigen). The studies that recorded unscheduled physician or ER visits as an efficacy parameter reported mean values of 2 visits/patient/year (Octagam) to 5.6 visits/patient/year (Gammaplex).
Antibiotic administration was reported in four studies (Table
1). Therapeutic antibiotic days/patient/year were very similar for the three studies that reported it. Prophylactic antibiotics were reported as 81 days/patient/year in the Flebogamma 5 % study and 46 days/patient/year in the Flebogamma 10 % study. Two studies reported only the total number of antibiotic days/patient/year—Privigen (87.4 days) and Biotest-IVIG (82 days). Total antibiotic days/patient/year in the Flebogamma 5 % and Flebogamma 10 % studies were 136.7 and 103.8, respectively.
IgG half-life values ranged from 20 ± 4 days in the Biotest-IVIG 21-day infusion group to 37 ± 13 days in the Flebogamma 10 % 28-day infusion group. IgG half-lives reported without segregation into infusion groups ranged from 34 to 41 ± 17 days. When recorded, the half-lives of IgG subclasses and specific antibodies were the essentially the same as the total IgG. The lack of standardized test procedures for determining specific antibody concentrations renders the comparison of antibody trough and other pharmacokinetic parameters difficult.
The reporting of AEs varied from study to study, with temporally associated AEs varying from 30 min to 72 h post-infusion. Using the 72-h time frame, the incidence of infusions with one or more temporal AEs (regardless of causality) produced AE rates of 20.1–47 % (Table
2). AEs considered to be related to the study IVIG ranged from 5 % of infusions for Octagam to 27.6 % of Flebogamma 10 % infusions (Table
2). Headache was the most common AE in every study.
The appearance of Coombs positivity after infusion of the study IVIGs occurred in four clinical trials. The proportion of patients that became Coombs-positive after infusion varied from 8.5 to 47 %. However, none of the patients in any of the trials developed evidence of hemolysis or anemia.
Clinical results with the newest IVIG from Biotest compared favorably to the results obtained with other IVIGs. Investigators observed a relatively low incidence of non-serious infections (2.6 per patient per year), few missed days of work/school (2.28 per patient per year), and a low rate of hospitalizations (0.21 per patient per year). The number of days of antibiotic treatment during the Biotest-IVIG study (82 days) were also lower than reports from other studies (range 87.4–136.7 days). AE rates were comparable to the rates reported for other products.
In summary, a review of IVIG clinical trial results reported in the new millennium show that clinical trial protocols are becoming more and more standardized, thanks to efforts of the FDA. Although comparisons should be conducted cautiously, the standardization of clinical trial data offers the opportunity to compare results from study to study. If standardized tests for quantifying specific antibodies are developed, pharmacokinetic studies may provide more meaningful information with respect to appropriate IVIG doses to prevent infection.
Acknowledgments
The technical support provided by John Hooper (BioCatalyst LLC, Liberty, Missouri, USA) is greatfully appreciated.