050
`
`American Journal of Hematology 34:90-98 (1990)
`
`Diagnosis of Cobalamin Deficiency I: Usefulness of
`Serum Methylmalonic Acid and Total
`Homocysteine Concentrations
`
`Robert H. Allen, Sally P. Stabler, David G. Savage, and John Lindenbaum
`Department of Medicine and Department of Biochemistry, Biophysics and Genetics, University of Colorado Health Sciences Center,
`Denver (R.H.A., S.P.S.); Department of Medicine, College of Physicians and Surgeons, Columbia University and Department of
`Medicine, Columbia-Presbyterian and Harlem Hospital Centers, New York (D.G.S., J.L.)
`
`The serum cobalamin assay is the primary diagnostic test for cobalamin deficiency. It
`appears to be an excellent screening test since most patients with clinically confirmed
`cobalamin deficiency have low levels. Recent studies indicate that the clinical picture of
`cobalamin deficiency is much more diverse than previously believed. It is also apparent
`that many patients with low serum cobalamin concentrations are not cobalamin deficient.
`Thus, there is a need for additional diagnostic tests to further distinguish patients with
`low serum cobalamin levels who are actually cobalamin deficient and will benefit from
`lifetime treatment from those who are not deficient and will not benefit. Serum levels of
`methylmalonic acid and total homocysteine have been shown to be markedly elevated in
`most patients with cobalamin deficiency, and total homocysteine concentrations are
`markedly elevated in most patients with folate deficiency. The levels of these metabolites
`fall to normal if these patients are treated with the appropriate vitamin but remain essen(cid:173)
`tially unchanged if the wrong vitamin is administered. These observations demonstrate
`that serum methylmalonic acid and total homocysteine levels are useful in diagnosing
`patients with cobalamin and folate deficiency and in distinguishing between these two
`vitamin deficiencies.
`
`Key words: methylmalonic acid, homocysteine, cobalamin, vitamin 8 12, folate
`
`INTRODUCTION
`
`Although most noted for the cobalamin (vitamin B 12 ,
`Cbl) absorption test that bears his name, Robert F.
`Schilling, M.D. has made other major contributions re(cid:173)
`lated to the diagnosis of Cbl deficiency. His prominence
`in this area is illustrated by the fact that in 1979 when the
`Food and Drug Administration asked the National Com(cid:173)
`mittee for Clinical Laboratory Standards (NCCLS) to
`appoint a panel to investigate major problems in serum
`Cb! assays l 1-3), Dr. Schilling was chosen to chair this
`panel. One of these problems related to the observation
`that previously unknown Cbl analogues were present in
`human blood. Scrum Cb! assays in use at that time con(cid:173)
`tained nonspecific Cbl-binding proteins that bound and
`measured the Cbl analogues together with Cbl, and thus
`gave erroneously high results. This finding was of diag(cid:173)
`nostic importance because approximately 20% of pa(cid:173)
`tients with clinically confirmed severe Cb! deficiency
`had values for the total of Cb! and Cb! analogues that
`
`© 1990 Wiley-Liss, Inc.
`
`were within the normal range, as shown in Figure l.
`When the nonspecific R protein was replaced with the
`specific intrinsic factor, only Cbl was measured, and
`lower values were obtained. In addition, and most im(cid:173)
`portantly, a complete separation between the normal sub(cid:173)
`jects and those with severe clinically confirmed Cbl de(cid:173)
`ficiency was obtained (Fig. 1 ).
`It is of interest that Dr. Schilling was the first to pub(cid:173)
`lish about differences in specificity of Cbl binding pro(cid:173)
`teins when in 1956 he wrote [ 4],
`
`Address reprint requests to Robert H. Allen, M.D., Division of He(cid:173)
`matology, Campus Box B 170, University of Colorado Health Sci(cid:173)
`ences Center, 4200 E. Ninth Avenue, Denver, CO 80262.
`
`This work was supported by Department of Health and Human Ser(cid:173)
`vices Research Grants (DK3 l 765 and DK2 l 365) from the National
`Institute of Diabetes and Digestive and Kidney Diseases.
`
`Teva – Fresenius
`Exhibit 1017-00001
`
`

`
`NORMAL SUBJECTS (74) o ; PATIENTS (21) •
`
`Cbl
`+
`Analogue
`(R)
`--~to--
`
`00
`0
`
`-~o:!E~-
`.
`a• ••
`....
`..
`='·
`
`Analogue
`IR minus IF)
`
`.
`
`---i:JO.....--
`
`- - -~-- •
`-*-
`
`Cbl
`(IF)
`
`0
`
`'-QO
`--~--
`
`..
`..
`•i· ...
`
`1300
`
`{ 700
`en
`~
`c
`E 300
`_Q
`0
`.D
`0
`u
`E 100
`:> .x
`
`30
`
`10 ..__ ______________ _
`
`Fig. 1. Serum Cbl values for 74 normal subjects (o) and 21
`patients with clinical evidence of Cbl deficiency (•) deter(cid:173)
`mined by radioisotope dilution assays using various Cbl(cid:173)
`binding proteins. Solid and dashed lines represent the
`mean ± 2 S.D. for the 74 normals. Cbl analogue levels were
`obtained by subtraction of values obtained with human in(cid:173)
`trinsic factor (IF) from those obtained with human salivary R
`protein (R) [adapted from reference 2 with permission].
`
`The in vitro binding of cyanocobalamin by gastric
`juice is a selective process showing a distinctive prefer(cid:173)
`ence for cyanocobalamin over pseudo vitamin B 12 [a Cbl
`analogue that contains adenine in place of 5,6-dimethyl(cid:173)
`benzimidazole] ...
`
`The process of cobalamin binding by serum, unlike that
`in gastric juice, does not manifest a selectivity for cy(cid:173)
`anocobalamin in the presence of excess pseudo vitamin
`
`B12·
`
`A second problem with the Cb! assays in use in the
`mid- l 970s involved imprecision with respect to estab(cid:173)
`lishing nonnal ranges. The two largest manufacturers of
`Cb! assay kits stated that 200 pg/ml was the appropriate
`lower limit of normal for their nonspecific Cb! assays
`[2]. Based on the data in Figure 1, together with data
`from other manufacturers [2] and subsequent investiga(cid:173)
`tors [5,6], it is clear that the appropriate lower limit of
`normal for the nonspecific Cb! assays is close to 300
`pg/ml. The use of a lower limit of normal of 200 pg/ml
`for the nonspecific Cbl assays resulted in ''normal val(cid:173)
`ues" being obtained for about 50% of severe clinically
`confirmed Chi-deficient patients (see Fig. 1).
`The NCCLS panel chaired by Dr. Schilling made a
`number of recommendations that included the use of spe(cid:173)
`cific binding proteins in Cbl assays and the recommen-
`
`Diagnosis of Cobalamin Deficiency I
`
`91
`
`dation that every Cbl assay be verified using test results
`obtained with normal subjects and clinically confirmed
`Chi-deficient patients. These recommendations have
`now been met for essentially all Cbl assays. As a result,
`package inserts for Cbl assays contain more clinical doc(cid:173)
`umentation than is found with virtually any other clinical
`assay.
`Although the diagnostic sensitivity of Cb! assays was
`markedly improved by the adoption of the NCCLS rec(cid:173)
`ommendations, problems were soon noted with respect
`to their diagnostic specificity, especially if they were
`used to screen apparently healthy individuals for Cb!
`deficiency. This problem was explicitly stated by Dr.
`Schilling in 1982 when he wrote [7],
`
`As a result of the change in observed levels of the
`vitamin with the newer radioligand kits, it is highly prob(cid:173)
`able that physicians will soon be required to explain in(cid:173)
`creasing numbers of unexpectedly low serum vitamin
`B 12 concentrations in patients with no other signs of
`deficiency.
`
`The potential enormity of this problem is apparent when
`one realizes that by definition 2.5% of normal healthy
`subjects will have low serum Cbl levels. Based on a
`population of 250,000,000, one can calculate that there
`are approximately 6,250,000 people in the United States
`with low serum Cbl values. It has been estimated that
`between 0.1 % and 1.0% of the population will become
`Cb! deficient at some time in their life [8]. If one assumes
`an average lifetime of 75 years and that Cb! deficiency
`will exist for an average of 5 years before it is treated or
`the patient dies, one can calculate that at any point in
`time there are 15,000 to 150,000 Chi-deficient patients
`in the United States. Even if all of these patients have
`low serum Cb! values, which seems unlikely, they will
`represent only 0.24% to 2.4% of individuals with low
`serum Cb! values.
`The magnitude of the difference between the incidence
`of low serum Cb! values and that of Cb! deficiency sug(cid:173)
`gests that problems will still remain even if one limits
`Cb! testing to those individuals who have hematologic or
`neuropsychiatric abnormalities of the kind caused by Cb!
`deficiency [9]. This is particularly likely because none of
`these abnormalities is specific for Cbl deficiency and
`because recent studies [10-13] have shown that the clin(cid:173)
`ical picture of Cb! deficiency is much more diverse than
`previously believed. Thus, many Cbl-deficient patients
`are not anemic, have normal mean cell volumes, normal
`white blood cell and platelet counts, normal peripheral
`smears as reported by routine laboratories, and normal
`lactate dehydrogenase and bilirubin levels. In addition,
`many patients with neuropsychiatric abnormalities that
`respond to Cb! therapy lack most, or even all, of the
`
`Teva – Fresenius
`Exhibit 1017-00002
`
`

`
`92
`
`Allen et al.
`
`COOH
`
`yOOH
`HG CH3
`
`CO S CoA
`
`D-Methylmalonyl-CoA
`
`Adenosyl-Cbl
`
`CO-S-CoA
`
`Mutase
`
`L -Methylmalonyl-CoA
`
`COOH
`yH
`
`2
`
`?H2
`
`CO-S-CoA
`Succinyl-CoA
`
`CH3-Tetrahydrofolate
`
`T etr ahydrofolate
`
`Methyl-Cb!
`Synthetase
`
`~ ~ ~H2
`HS-C-C-C-COOH
`
`H H H
`
`~ ~ ~H2
`CH 3 -S-y-9-9 COOH
`H H H
`
`Homocysteine
`
`Methionine
`
`Fig. 2. L-methylmalonyl-CoA mutase (left) and methionine synthetase (right) are the two
`mammalian Cbl-dependent enzymes. In Cbl deficiency, L-methylmalonyl-CoA accumu(cid:173)
`lates and is converted to D-methylmalonyl-CoA, which is hydrolyzed to methylmalonic
`acid. Homocysteine accumulates in either Cbl or folate deficiency.
`
`hematologic abnormalities [ 13]. One cannot solve this
`problem by using a value of 100 pg/ml as the cut-off
`between Cbl deficiency and normality since the same
`recent studies [ 10-13] have shown that approximately
`40% of clinically confirmed Cbl-deficient patients have
`serum Cbl values in the JOO to 200 pg/ml range.
`In the early 1980s, the additional diagnostic tests for
`Cbl deficiency that were widely available were limited to
`treating patients with Cbl and looking for objective re(cid:173)
`sponses to such therapy or to use of the Schilling test.
`Therapeutic trials may require months before responses
`can be evaluated and can be difficult to interpret in pa(cid:173)
`tients with neuropsychiatric abnormalities since these do
`not always respond to Cbl therapy even if Cb! deficiency
`is the cause of the abnormalities. The use of the Schilling
`test as a diagnostic test for Cbl deficiency has been crit(cid:173)
`icized by Dr. Schilling who wrote in 1982 [7],
`
`A technique often mistakenly utilized as a means for
`establishing or refuting vitamin B 12 deficiency is mea(cid:173)
`surement of the patient's ability to absorb an orally ad(cid:173)
`ministered dose of the vitamin. Such estimation of the
`ability to absorb vitamin B 12 may be helpful in delineat(cid:173)
`ing the mechanism by which a patient has become B 12
`deficient or in predicting which patients might become
`deficient in the vitamin while on a normal diet, but test(cid:173)
`ing the patient's ability to absorb the vitamin will not
`give the diagnosis of deficiency per se.
`
`Nevertheless, as a practical point, few would argue with
`the widely accepted practice of instituting lifetime Cb!
`therapy in patients with abnormal Schilling tests. Al(cid:173)
`though some patients are unable to collect proper 24 hour
`urine specimens as required in the Schilling test [ 14], the
`major problem with using it as the sole additional diag(cid:173)
`nostic test in patients with low serum Cb! values is that
`
`the classic Schilling test uses crystalline Cb!, whereas
`Cb! in food is tightly bound to protein and this protein(cid:173)
`bound Cbl is what must be absorbed under physiologic
`conditions. Studies by Doscherholmn and Swaim [15]
`and others [ 16, 17] have documented that patients with
`hypochlorhydria are unable to absorb food Cb!, even
`though their ability to absorb crystalline Cb! remains
`intact. Furthermore, recent studies by Carmel et al. [ 18]
`have shown that food Cbl malabsorption occurs com(cid:173)
`monly in patients with low serum Cbl levels.
`It is of interest that studies [ 19] by Dr. Schilling in
`1967 anticipated the importance of food Cbl malabsorp(cid:173)
`tion. Dr. Schilling injected radioactive Cbl into chickens
`and obtained eggs that contained radioactive Cb!. He
`showed that patients with acid gastric juice containing
`pepsin absorbed more of the egg Cb! than did pernicious
`anemia patients supplemented with oral intrinsic factor.
`He suggested that '' ... peptic activity is important in
`food B 12 absorption."
`In the early 1980s, we set out to develop new diag(cid:173)
`nostic tests for Cb! deficiency per se with an emphasis on
`assays that could be performed with serum since this is
`what is left in the laboratory after serum Cb! is measured.
`Because Cb! serves as a co-factor for only two mamma(cid:173)
`lian enzymes, we concentrated on compounds related to
`these two pathways, which are shown in Figure 2. Al(cid:173)
`though many compounds were tested using gas chroma(cid:173)
`tography-mass spectrometry
`[20,21], methylmalonic
`acid and total homocysteine were the only ones with
`diagnostic utility [22,23]. The utility of measuring serum
`methylmalonic acid was not surprising since many stud(cid:173)
`ies dating from the 1950s have shown that many Cbl(cid:173)
`deficient patients excrete increased amounts of methyl(cid:173)
`malonic acid in their urine [24,25].
`The initial studies [22,23) of normal subjects and pa(cid:173)
`tients with moderate and severe clinically confirmed Cb!
`
`Teva – Fresenius
`Exhibit 1017-00003
`
`

`
`30,000
`
`3000
`
`1000
`
`~ 10,000
`......
`Ol s
`"O
`'ti
`<(
`u
`·c:
`0
`~
`E
`>.
`..c
`Qi
`~ 100
`E
`::J
`Qi
`Cf)
`
`300
`
`30
`
`.,
`..
`et) .. -.e
`~ • I • j
`• ..
`__ .. -
`-.91-,
`...
`
`-.!<..-
`... ~
`-~
`--.- -
`·::
`
`10
`
`Normals
`(n=50)
`
`Cbl-deficient
`(n = 7 3)
`
`Fig. 3. Values for serum methylmalonic acid obtained with
`normal subjects and patients with clinically confirmed Cbl
`deficiency. Solid and dashed lines represent the mean ± 2
`S.D. for the normal subjects [adapted from reference 22 with
`permission].
`
`deficiency are shown in Figures 3 and 4. Approximately
`95% of these Chi-deficient patients had elevations of
`serum methylmalonic acid, and the same was true for
`serum total homocysteine. In recent studies [11] with 86
`consecutive Cbl-deficient patients with serum Cbl levels
`<200 pg/ml, many of whom had mild Cb! deficiency
`although all had objective responses to Cbl therapy, 77%
`had marked elevations (>3 S.D. above the mean for
`normal subjects) of both serum methylmalonic acid and
`homocysteine. Another 9% had a marked elevation of
`methylmalonic acid alone, and 8% had a marked eleva(cid:173)
`tion of total homocysteine alone. Only 6% failed to have
`a marked elevation of either metabolite. Based on these
`results, we believe that measurement of both serum meth(cid:173)
`ylmalonic acid and total homocysteine is often required
`for the optimal diagnosis of Cbl deficiency.
`Serum levels of methylmalonic acid and total homo(cid:173)
`cysteine are particularly useful in patients with neuropsy(cid:173)
`chiatric abnormalities due to Cb 1 deficiency who lack
`anemia or macrocytosis, since all such patients evaluated
`in a recent study had a marked elevation of at least one
`metabolite [ 13].
`About 90% of patients with clinically confirmed folate
`deficiency have markedly elevated levels of serum total
`homocysteine [23]. Our initial studies indicated that
`about 20% of folate-deficient patients also had increases
`in serum methylmalonic acid, but re-evaluation of these
`patients indicates that most of these elevations were due
`
`Diagnosis of Cobalamin Deficiency I
`
`93
`
`400
`
`':J
`......
`0
`E
`3 100
`Q)
`c
`·a;
`©
`>.
`()
`0
`E
`0
`J:
`
`40
`
`10
`
`..
`:a
`••;£\•••
`···~·
`:~.tc:i-•.
`..
`··~
`..
`:-
`.:
`
`...
`.
`-.
`..
`
`;:
`I
`
`~~~~
`--~·--
`
`4
`
`Normals
`(n=50)
`
`Cbl-deficient
`(n = 78)
`
`Folate-
`deficient
`(n =19)
`
`Fig. 4. Values for serum total homocysteine obtained with
`normal subjects, patients with clinically confirmed Cbl de(cid:173)
`ficiency, and patients with clinically confirmed folate defi(cid:173)
`ciency. Solid and dashed lines represent the mean ± 2 S.D.
`for the normal subjects [adapted from reference 23 with per(cid:173)
`mission].
`
`to renal insufficiency or intravascular volume contrac(cid:173)
`tion. Less than 2% of folate-deficient patients with nor(cid:173)
`mal renal function and volume status have elevated lev(cid:173)
`els of serum methylmalonic acid, and these elevations
`are modest ( <500 nmoles per liter). These results indi(cid:173)
`cate that measurement of both serum methylmalonic acid
`and total homocysteine can differentiate between Cbl and
`folate deficiency in most cases.
`Serum methylmalonic acid and total homocysteine
`levels also provide useful information when they are ob(cid:173)
`tained both before and after therapy with Cb! or folate
`[22,23]. As shown in Figure 5, serum methylmalonic
`acid and total homocysteine values fell toward normal
`within several days of treatment with Cb! in a patient
`with pernicious anemia. Levels of serum total homocys(cid:173)
`teine fell in a similar manner (Fig. 6) after folate treat(cid:173)
`ment in a patient with folate deficiency.
`The availability of assays for serum methylmalonic
`acid and total homocysteine has made it possible to per(cid:173)
`form a number of additional investigations in patients
`with Cb! and folate deficiency. These studies include the
`following: l) The response of hematologic findings and
`serum levels of methylmalonic acid and total homocys(cid:173)
`teine in patients with Cb! or folate deficiency who were
`treated with the incorrect vitamin; 2) estimation of the
`prevalence of patients with normal Cbl levels and ele(cid:173)
`vated methylmalonic acid and total homocysteine levels
`who had subsequent objective responses to Cb! therapy;
`3) estimation of the relative sensitivities of serum Cbl,
`methylmalonic acid, and total homocysteine levels in
`previously diagnosed Chi-deficient patients who have
`been without parenteral Cb! therapy for intervals of 2 to
`66 months; and 4) the influence of the gut flora on levels
`
`Teva – Fresenius
`Exhibit 1017-00004
`
`

`
`94
`
`Allen et al.
`
`1000
`
`:Q 300
`u
`<(
`u ·c
`0
`lii 100
`E
`>.
`.c a;
`
`~
`
`30
`
`-!Normal -
`Range
`
`. . - - - .
`
`100
`
`......
`
`~
`Q)
`
`--..
`0
`E 30
`
`E
`0 :r:
`E
`2
`
`Q)
`Cf)
`
`3
`
`3 l 10 I:~~:' ___ _
`
`0
`
`5
`
`10
`5
`15
`0
`Days after Treatment with Parenteral Cb/
`Fig. 5. Levels of serum methylmalonic acid and total homocysteine in a patient with Cbl
`deficiency due to pernicious anemia before and after treatment with parenteral Cbl.
`Dashed lines represent 2 S.D. above and below the mean for normal subjects [adapted
`from references 22 and 23 with permission].
`
`10
`
`15
`
`50
`
`?
`0
`E 40
`3
`Q) c
`<ii
`c;; 30
`>.
`(.)
`0
`E
`0
`I
`(ii
`0
`I-
`E 10
`:l
`Qj
`(/)
`
`20
`
`0
`
`l Normal
`
`Range
`
`-1
`7
`2
`5
`4
`3
`6
`0
`Days After Treatment with Oral Folic Acid ( 1mg/day)
`
`Fig. 6. Levels of serum total homocysteine in a patient with
`folate deficiency due to alcoholism before and after treat(cid:173)
`ment with oral folic acid. Dashed lines represent 2 S.D.
`above and below the mean for normal subjects [adapted
`from reference 23 with permission].
`
`of methylmalonic acid and total homocysteine, which
`was evaluated by studying Cbl-deficient patients treated
`with antibiotics. The remainder of this report contains
`the results of the first set of studies. The latter three are
`contained in the accompanying paper in this series [26].
`
`MATERIALS AND METHODS
`
`binding protein (Quantiphase, BioRad Laboratories,
`Richmond, CA). Methylmalonic acid and total homocys(cid:173)
`teine concentrations were measured by modifications of
`recently developed techniques [21,22,27] using capil(cid:173)
`lary-gas chromatography and mass spectrometry . 1 The
`major modifications, which will be described in detail
`elsewhere, consist of the following: I) the HPLC step in
`the methylmalonic acid assay was replaced by a step that
`involved the adsorption and elution from a disposable
`silica column or a single anion-exchange column; 2) ho(cid:173)
`mocysteine was reduced with dithiothreitol, treated with
`iodoacetamide, and further purified using a single cation
`exchange column; and 3) the rate of temperature increase
`during gas chromatography was increased from 8°C per
`minute to 30°C per minute for both assays.
`The within-run precision (CV) was approximately 2%,
`and the between-run precision was approximately 6% for
`both modified assays. Revised normal ranges were ob(cid:173)
`tained with the modified assays using 50 normal human
`blood donors, 25 males and 25 females, who ranged in
`age from 18 to 65 years. Normal ranges were calculated
`as the mean :±: 2 S.D. after log normalization to correct
`for the skewness of data toward higher values. The re(cid:173)
`vised normal range for serum methylmalonic acid was
`73-271 nmoles/liter and for serum total homocysteine
`was 5.4-16.2 µ,moles/liter. The corresponding ranges
`for the mean :±: 3 S.D. were 53-376 nmoles/liter and
`4.1-21.3 µ,moles/liter.
`
`Serum Cbl and folate levels were determined by ra(cid:173)
`dioassay using purified intrinsic factor and purified fol ate
`
`1 Aspects of these assays are the subject of patent applications filed on
`behalf of the University of Colorado and Columbia University.
`
`Teva – Fresenius
`Exhibit 1017-00005
`
`

`
`40,000
`
`20,000
`
`_J
`
`0
`
`2,000
`
`~ 10,000
`E
`:i
`"O ·u 5,000
`ro
`(.) c
`ro
`E
`:;,
`..c
`Ql
`E
`E
`::::l
`(ii
`(f)
`
`1,000
`
`500
`
`Diagnosis of Cobalamin Deficiency I
`
`95
`
`110
`
`_J
`-...
`0 90
`E
`:i
`(l)
`c
`'Qi
`(i)
`>. g 60
`E
`0
`..c
`ro
`§
`E
`2 30
`(l)
`(j)
`
`-e
`
`200
`
`100
`
`I FOLIC ACID I
`
`r -1 -C-0-BA_L_A-Ml-N--,I
`
`Before
`
`After
`Therapy
`
`Atter
`
`Fig. 7. Serum methylmalonic acid concentrations in 5 pa(cid:173)
`tients with Cbl deficiency who were first treated with folic
`acid only and later with CN-Cbl only. The duration of folate
`therapy prior to repeat assays of methylmalonic acid was
`12, 13, 23, 65 and 67 days for patients a, b, c, d and e,
`respectively. See Results for additional details concerning
`these patients and their treatment. Dashed line indicates 3
`S.D. above the mean value for normal subjects using the
`revised assay as described under "Materials and Methods."
`
`Before
`
`Atter
`Therapy
`
`Atter
`
`Fig. 8. Serum total homocysteine concentrations in 5 pa(cid:173)
`tients with Cbl deficiency who were first treated with folic
`acid only and later with CN-Cbl only. The duration of folate
`therapy prior to repeat assays of total homocysteine was 12,
`13, 23, 65 and 67 days for patients a, b, c, d and e, respec(cid:173)
`tively. See Results for additional details concerning these
`patients and their treatment. Dashed line indicates 3 S.D.
`above the mean value for normal subjects using the revised
`assay as described under "Materials and Methods."
`
`RESULTS AND DISCUSSION
`Inappropriate Therapy of Cbl Deficiency
`With Folate
`It is well known that although inappropriate treatment
`of Cb! deficiency with large doses of folic acid may often
`result in hematologic response, deterioration of neuro(cid:173)
`logic function may occur [28,29]. If normalization of
`elevated serum metabolite levels after Cb! therapy is to
`be used as a criterion for the presence of Cb! deficiency
`[22,23], then it is important to know the effects of folic
`acid treatment alone on serum methylmalonic acid and
`total homocysteine concentrations in such Chi-deficient
`patients.
`Data are available from nine patients, all of whom had
`underlying Cb! deficiency without evidence of folate de(cid:173)
`ficiency, who were treated with varying doses of folic
`acid. In the first five patients, serum was available prior
`to and immediately after folate therapy as well as fol(cid:173)
`lowing subsequent treatment with Cb! (Figs. 7, 8). Four
`of them had well-documented pernicious anemia, and the
`fifth (patient b, Figs. 7, 8) had biopsy-proven tropical
`
`sprue. Prior to therapy, all had low serum Cb! concen(cid:173)
`trations and normal or elevated serum folate levels. Pa(cid:173)
`tients a, b, and c had macrocytic anemias associated with
`megaloblastic bone marrow morphology. In patients d
`and e, the hematocrit was normal, although hyperseg(cid:173)
`mented neutrophils and macroovalocytes were present on
`blood smears and the MCVs were 109 and 98 fl, respec(cid:173)
`tively. Patients a and b were treated with serial regimens
`of low vitamin doses as part of therapeutic trials. They
`received 150-200 µg of folic acid intramuscularly daily
`for 10 to 13 days without reticulocyte or other hemato(cid:173)
`logic responses. They subsequently responded fully to
`2-week trials of 1-5 µg daily of intramuscular CN-Cbl.
`Patients c, d, and e were given folic acid owing to phy(cid:173)
`sician error. Patient c received 1 mg of folic acid by
`mouth for 23 days without any change in the reticulocyte
`count while the hematocrit fell; she subsequently re(cid:173)
`sponded fully to 7 days of 1 mg parenteral CN-Cbl.
`Patients d and e were given 1 mg folic acid daily by
`mouth for 2 to 2.5 months; both had hematologic re(cid:173)
`sponses. In patient d, the red cell count rose from 3.68 x
`106 /µI to 4.57 x 106 /µI, the MCV fell from 106 to 89
`
`Teva – Fresenius
`Exhibit 1017-00006
`
`

`
`96
`
`Allen et al.
`
`TABLE I. Additional Cbl-Deficient Patients Treated With Folic Acid Alone*
`
`Immediately after completion of folic acid treatment
`
`Patient
`
`Diagnosis·'
`
`Duration
`(mon)
`
`g
`h
`
`IR
`PA
`PA
`
`JD
`
`3
`17
`2
`6
`18
`
`Daily
`dose
`(mg)
`
`0.4
`
`Neurologic
`deterioration
`+
`+
`?
`
`Hematologic
`response
`+
`+
`+
`+
`0
`
`Serum
`Cbl
`(pg/ml)
`
`Serum
`folate
`(ng/ml)
`
`Serum
`methylmalonic
`acid
`(nmol/L)b
`
`Serum
`total
`homocysteine
`(µmol/L)'
`
`80
`150
`140
`80
`65
`
`>50.0
`23.0
`14.5
`15.4
`>50.0
`
`241,000
`26,380
`2,040
`3,350
`910
`
`77
`200
`45
`37
`89
`
`+
`*Patients f and g were reported previously ([ref. 13] patients 24 and 30, respectively).
`"IR 0= ilea! resection; PA = pernicious anemia (intrinsic factor antibodies present in serum); JD = jejuna! diverticulosis.
`"Normal range (based on mean ± 3 S.D.) = 53-376 nmol/L.
`'Normal range (based on mean± 3 S.D.) = 4.l-21.3 umol/L.
`
`fl, while the hematocrit remained unchanged at 39%. In
`patient 3, the hematocrit remained stable at 47%, but the
`MCV fell from 98 to 94 fl and the blood smear became
`normal after folate treatment. Subsequently, patients d
`and e received 3 injections of I mg cyanocobalamin.
`None of these patients experienced a change in neuro(cid:173)
`logic symptoms during folate treatment.
`After folic acid therapy, the serum methylmalonic acid
`level remained clearly elevated in all five patients (Fig.
`7). In four the level rose, and in another it showed a
`modest decline. In contrast, in all five patients the serum
`methylmalonic acid fell quickly into the normal range
`after the institution of Cb! treatment. Similarly, the se(cid:173)
`rum total homocysteine concentration (Fig. 8), although
`it rose somewhat in three and fell somewhat in two pa(cid:173)
`tients, was clearly elevated after folic acid therapy in all
`five patients and fell promptly into the normal range after
`Cb! was given.
`In an additional four patients with Cbl deficiency (Ta(cid:173)
`ble I), prolonged courses of folic acid, 0.4 to I mg daily
`by mouth, were given (owing to physician error) result(cid:173)
`ing in hematologic responses in three. In patient f, the
`hemoglobin rose from 11.4 to 13.2 g/dl, but progressive
`visual loss due to optic atrophy and changes in person(cid:173)
`ality occurred while on folate. In patient g, who had
`alpha-thalassemia trait, the hematocrit rose from 25% to
`369C, and the MCV fell from 93 to 79 fl during folic acid
`treatment although paresthesias and sensory deficits in
`the hands and feet progressed. In patient h, the hemat(cid:173)
`ocrit rose from 35% to 46% and the MCV decreased
`from 105 to 82 fl while her position sense abnormalities
`were unchanged; dementia progressed on folate but
`showed no subsequent response to Cbl. Patient i did not
`respond hematologically but developed ataxia after 15
`months on folic acid. Unfortunately, baseline serum
`prior to folic acid treatment was not available in these
`four patients. Metabolite levels during folate therapy are
`shown in Table I. In patients f and g, methylmalonic acid
`
`concentrations were well above the median (3,400 nmol/
`L) in a series of 351 patients with clinical Cb! deficiency
`that we have studied (unpublished data). Similarly, the
`serum total homocysteine levels in patients g and i ex(cid:173)
`ceeded the median (72 µmol/L) in the series of 351 pa(cid:173)
`tients. After Cb! therapy, serum metabolite levels be(cid:173)
`came normal in the three patients (g, h, and i) from
`whom follow-up serum was available.
`In summary then, in all nine patients with Cb! defi(cid:173)
`ciency treated with folic acid, including five who re(cid:173)
`sponded hematologically, the serum metabolite levels re(cid:173)
`mained clearly elevated or even rose further.
`
`Inappropriate Therapy of Folate Deficiency
`With Cbl
`At present, data are available on three patients with
`folate deficiency who were treated with CN-Cbl alone
`(Table II). In all three, the serum total homocysteine was
`elevated before therapy, and the serum methylmalonic
`acid was within normal limits. The Schilling test, per(cid:173)
`formed at the beginning of CN-Cbl treatment, was nor(cid:173)
`mal in each. The timing of follow-up data in these pa(cid:173)
`tients was such
`that
`it was uncertain whether a
`hematologic response to Cbl treatment occurred. How(cid:173)
`ever, the serum total homocysteine concentration re(cid:173)
`mained elevated after Cb! therapy in each patient (Table
`II). The serum level returned to normal following folic
`acid treatment in all three patients. Although patient C
`received only one injection of CN-Cbl, in our experience
`to date the lack of any change in the serum total homo(cid:173)
`cysteine has not been encountered in a patient with Cbl
`deficiency. In six patients with Cb! deficiency that we
`have studied during the l st week after treatment with a
`single injection of l mg of CN-Cbl, the serum homocys(cid:173)
`teine fell in every patient, declining to less than 60% of
`the baseline value in 5. These observations thus suggest
`that elevated total homocysteine levels secondary to fo(cid:173)
`late deficiency do not respond to Cbl therapy.
`
`Teva – Fresenius
`Exhibit 1017-00007
`
`

`
`TABLE II. Folate-deticient Patients Treated With CN-Cbl Alone
`
`Patient
`
`Age (years)lsex
`
`Cause of
`fol ate
`deficiency
`
`Serum
`
`Hematocrit MCV
`(fl)
`(%)
`
`Cb!
`(pg/ml)
`
`Folate
`(nglml)
`
`Diagnosis of Cobalamin Deficiency I
`
`97
`
`CN-Cbl
`treatment
`I mg x 3 over I mo.
`I mg x 4 over 5 wks
`I mg
`
`Serum total homocysteine
`(µmol/L)"
`
`Untreated
`
`After
`Cb lb
`
`After
`folate
`
`39
`73
`29
`
`36
`59
`52
`
`18
`21
`6
`
`114
`124
`108
`
`225
`140
`140
`
`2.4
`l.6
`l.6
`
`77/F
`68/F
`36/F
`
`?Alcoholic
`?Triamterene
`Dietary
`
`40
`41
`37
`
`A
`B
`c
`"Normal range (based on mean ± 3 S.D.) = 4.1-21.3 µmol/L.
`"Serum obtained one week after last injection of CN-Cbl.
`
`Patients in whom serum levels of both folate and Cbl
`are low are frequently encountered clinically and often
`pose a diagnostic problem. Although in more than 75%
`of patients with Cb! deficiency in our experience, serum
`levels of both methylmalonic acid and total homocys(cid:173)
`teine will be elevated, in about 10% only the homocys(cid:173)
`teine value is high. In addition, as shown in the accom(cid:173)
`this series [26], antibiotics may
`panying paper in
`normalize the serum methylmalonic acid level without
`affecting an elevated serum total homocysteine. Since an
`isolated elevation in total homocysteine is characteristic
`of folate deficiency [24], a therapeutic trial with a single
`vitamin, even in pharmacologic doses, may be useful in
`such patients, since our findings (Figs. 7, 8; Tables I, II)
`indicate that elevated metabolite levels will only fall to
`normal when therapy with the vitamin in which the pa(cid:173)
`tient is deficient is given.
`
`CONCLUSIONS
`
`Dr. Schilling has stated [7], "Because vitamin B 12
`deficiency may cause serious but eminently treatable he(cid:173)
`matologic and neurologic di