Letters

cluding insulin resistance and measures study (8), which assessed glucose over 1
of total and regional adiposity. COMMENTS AND year and compared clozapine and chlor-
Thus, these results show that hypo-
adiponectinemia is closely associated
RESPONSES promazine, are not mentioned in the bib-
liography.
with nonalcoholic hepatic steatosis in None of the retrospective studies can
obese healthy individuals, thus suggest- state how many patients given each drug
ing that hypoadiponectinemia might be, Consensus received blood tests. This is a crucial con-
at least partly, responsible for hepatic ste- Development founder, as patients receiving typical anti-
atosis and liver test abnormalities found psychotics have less blood monitoring
in these subjects. Interpretation of our re- Conference on than those receiving “atypicals” (9). Any
sults, however, requires care because of Antipsychotic Drugs study that introduces glucose screening
the relatively small number of patients. and Obesity and will undoubtedly find new previously
Future studies using larger cohorts will be undiagnosed diabetes because of the
needed to validate this hypothesis. Diabetes high prevalence of undiagnosed type 2
This clinical finding, however, is con- diabetes.
sistent with a recent study demonstrating Response to consensus statement I am unaware of any prospective trial
that adiponectin administration was ef- showing any difference among “atypicals”
fective in alleviating obesity-induced hep- in terms of emergent glucose abnormali-

I
atomegaly, steatosis, and serum ALT commend the authors of the consen- ties. The best trial data comparing aripi-
abnormality in mice (3). The potential sus statement in Diabetes Care for fo- prazole with olanzapine over 6 months
hepatoprotective mechanisms include in- cusing our attention on the physical found that emergent glucose abnormali-
duction of hepatic fatty acid oxidation, in- health of people with schizophrenia and ties were identical (4.7 vs. 4.5%) (5,7).
hibition of fatty acid synthesis, and serious mental illnesses and on the need The use of placebo cohorts is critical to
suppression of tumor necrosis factor-␣ for glucose monitoring (1). There is little understanding what part of the risk of
production in the liver. Therefore, in ad- consensus about the risk of diabetes in glucose abnormalities is attributable to
dition to its antidiabetic and antiathero- people with schizophrenia and the role of drugs. There are two such datasets, and
genic potentials, adiponectin or its atypical antipsychotic drugs (2– 4). Re- the incidence of diabetes in the placebo
agonists may represent a novel agent for cently I was a member of a group of inter- cohorts does not differ from that in the
the treatment of liver diseases. national psychiatrists and diabetologists active-drug group (5,6).
who reviewed the evidence surrounding Interestingly, during the randomized
GIOVANNI TARGHER, MD this issue (proceedings have been pub- control trials, weight gain was not associ-
LORENZO BERTOLINI, MD lished in a supplement to the April 2004 ated with the development of diabetes
LUCIANO ZENARI, MD issue of the British Journal of Psychiatry). I and the incidence of diabetes did not dif-
was, therefore, interested to see that the fer among the various “atypicals,” despite
conclusions published in Diabetes Care differing propensity for weight gain.
From the Diabetes Unit, Sacro Cuore Hospital, Ne- differed in some respects from our delib- Linking short-term weight gain to the risk
grar, Italy. erations and those of the U.S. Food and of diabetes ignores the many other genetic
Address correspondence to Giovanni Targher,
MD, Diabetes Unit, Ospedale “Sacro Cuore,” Via Drug Administration (FDA). In Septem- and environmental reasons why people
Sempreboni, 5, 37024 Negrar (VR), Italy. E-mail: ber 2003, the FDA wrote to the manufac- with schizophrenia develop diabetes
[email protected]. turers of all atypical antipsychotic drugs (10,11).
© 2004 by the American Diabetes Association. to ask that their respective labels be Antipsychotic medication is essential
changed to recommend regular glucose for people with schizophrenia, and effec-
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
testing for all schizophrenia patients at tiveness should be the most important
risk of diabetes. The FDA went on to ex- consideration when selecting treatment.
References
1. Angulo P: Nonalcoholic fatty liver disease.
plain that “[p]recise risk estimates for hy- The FDA was nearer to the mark in its
N Engl J Med 346:1221–1231, 2002 perglycemia-related adverse events in judgement, and choosing an antipsy-
2. Chandran M, Phillips SA, Ciaraldi T, patients treated with atypical antipsy- chotic drug on the basis of its potential to
Henry RR: Adiponectin: more than just chotics are not available.” worsen glycemia is failing to understand
another fat cell hormone? Diabetes Care The Diabetes Care consensus state- the available data.
26:2442–2450, 2003 ment suggests that the risk of hyperglyce-
3. Xu A, Wang Y, Keshaw H, Xu LY, Lam mia for some “atypicals,” such as
KSL, Cooper GJS: The fat-derived hor-
RICHARD I.G. HOLT, PHD, MRCP
clozapine and olanzapine, is greater than
mone adiponectin alleviates alcoholic and for others (1). It is unclear how this opin-
nonalcoholic fatty liver diseases in mice. From the Endocrinology and Metabolism Unit, De-
ion was reached, because although a bib- velopment Origins of Health and Disease Division,
J Clin Invest 112:91–100, 2003
4. López-Bermejo A, Botas P, Funahashi
liography was given, the work was School of Medicine, University of Southampton,
T, Delgado E, Kihara S, Ricart W, Fer- unreferenced. One may speculate that the Southampton, U.K.
weight given to the 35 retrospective stud- Address correspondence to Dr. R.I.G. Holt, Level
nández-Real JM: Adiponectin, hepato- F, Centre Block, MP 113, Southampton General
cellular dysfunction and insulin sensitiv- ies was greater than that for the prospec- Hospital, Tremona Road, Southampton SO16 6YD,
ity. Clin Endocrinol (Oxf) 60:256 –263, tive trials (4); the placebo-controlled U.K. E-mail: [email protected].
2004 studies (5–7) and the longest prospective R.I.G.H. has been a member of an advisory panel

2086 DIABETES CARE, VOLUME 27, NUMBER 8, AUGUST 2004

 L E T T E R S

for, has received honoraria from, and has received

grant support from Eli Lilly.
Consensus primacy of appropriate treatment, for ex-
ample with clozapine having unique ben-
© 2004 by the American Diabetes Association. Development efits for treatment-refractory patients and
Conference on those at significant risk for suicidal behav-
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● Antipsychotic Drugs ior, other antipsychotics may also have
a favorable efficacy profile among treat-
References and Obesity and ment-refractory patients in specific
1. American Diabetes Association: Consen-
sus development conference on antipsy-
Diabetes treatment domains such as cognitive
dysfunction (favoring risperidone and
chotic drugs and obesity (Consensus
olanzapine) (6). Efficacy considerations
Statement). Diabetes Care 27:596 – 601, Response to consensus statement are particularly important among patients
2004
2. Dinan TG, Kohen D: Atypical antipsy-
with persistent aggressive behavior
(again, favoring clozapine). Further com-

T
chotics and diabetic propensity: more he report from the Consensus De-
questions than answers. Hum Psychophar- velopment Conference on Antipsy- plicating the risk-benefit equation is the
macol Clin Exp 18:591–593, 2003 chotic Drugs and Obesity and observation that weight gain may be asso-
3. Buse JB: A retrospective cohort study of Diabetes Consensus Panel (1) contains ciated with response in about half of the
diabetes mellitus and antipsychotic treat- valuable advice for the appropriate and responders to clozapine or olanzapine (7).
ment in the United States. J Clin Epidemiol prudent monitoring of patients who are at In addition, interindividual differences in
56:164 –170, 2003 risk for type 2 diabetes. The recommen- response may be huge, leading to patients
4. Cerri K, Bushe CJ, Haddad P: How valid dations made are similar to a prior con- receiving several sequential medication
are retrospective epidemiological studies
sensus conference (2) and, if they are trials over the years to find the optimal
in evaluating differential rates of diabe-
followed, will benefit our vulnerable pa- regimen for that individual.
tes mellitus attributed to antipsychotics? As clinicians and researchers in a
Schizophr Res 67:S367, 2004 tients by the early identification and treat-
state-operated psychiatric center, where
5. Food and Drug Administration: Ablify ment of metabolic disorders.
the average patient has failed a number of
(aripiprazole) tablets [article online], However, the report probably over-
medication trials, we find obtaining an
2003. Available from http://www.fda.gov/ reaches available evidence when suggest-
adequate antipsychotic medication re-
cder/foi/nda/2002/21-436_Abilify.htm. ing that clinicians should consider
sponse to be a major challenge. When
Accessed 1 June 2004 prescribing one antipsychotic over an-
such a response is achieved, a major focus
6. Cavazzoni P, Mukhopadhyay N, Carlson other with the aim of avoiding diabetes.
is to manage somatic problems should
C, Brier A, Buse J: Retrospective analysis Although clear differences exist in liability they emerge. Efficacy is the prime mover
of risk factors in patients with treatment for weight gain (and consequently dys-
emergent diabetes during clinical trials of
for treatment decisions (8). A switch from
lipidemias), quantifiable risk differences a beneficial antipsychotic regimen is usu-
antipsychotic medications. Br J Psychiatry among the second-generation antipsy-
Suppl 47:S94 –S101, 2004 ally the last resort after other management
chotics regarding an association with dia- approaches have not been successful.
7. Bushe CJ, Leonard B: Association between betes have been inconsistent in large
atypical antipsychotic agents and type 2 Taken out of context, the consensus con-
published pharmacoepidemiological ference report might lead the inexperi-
diabetes: review of prospective clinical
studies (3). For this reason, the U.S. Food enced practitioner to switch medications
data. Br J Psychiatry Suppl 47:S87–S93,
2004 and Drug Administration has notified the prematurely and thus expose the patient
8. Lieberman JA, Phillips M, Gu H, Stroup S, manufacturers of the second-generation to the risk of a deterioration in their symp-
Zhang P, Kong L, Ji Z, Koch G, Hamer RM: antipsychotics that product labeling for toms and, ultimately, relapse.
Atypical and conventional antipsychotic all drugs in that class will require a new
drugs in treatment-naı̈ve first-episode warning about hyperglycemia and diabe- LESLIE CITROME, MD, MPH
schizophrenia: a 52-week randomized tes (4). Furthermore, the risk attributable JAN VOLAVKA, MD, PHD
trial of clozapine vs chlorpromazine. to antipsychotics appears small compared
From the Clinical Research and Evaluation Facility,
Neuropsychopharmacology 28:995–1003, with established risk factors such as fam- Nathan S. Kline Institute for Psychiatric Research,
2003 ily history and advancing age. From the New York University School of Medicine, New York,
9. Citrome L: Antipsychotic medication evidence, choosing a second-generation New York.
treatment and new prescriptions for in- antipsychotic medication does not, in and Address correspondence to Dr. Leslie Citrome,
sulin and oral hypoglycaemics. Eur Neu- New York University School of Medicine, Nathan S.
of itself, have significant predictive value Kline Institute for Psychiatric Research, Clinical Re-
ropsychopharmacol 13 (Suppl. 4):S306, for treatment-emergent diabetes. search and Evaluation Facility, Orangeburg, NY
2003
Moreover, the report did not ade- 10962. E-mail: [email protected].
10. Ryan MC, Collins P, Thakore JH: Im- L.C. has received honoraria from Abbott Labora-
paired fasting glucose tolerance in first-
quately address the complex issues regard-
tories, AstraZeneca, Bristol-Myers-Squibb, Eli Lilly,
episode, drug-naive patients with schizo- ing antipsychotic efficacy. Although Novartis, and Pfizer; has received research support
phrenia. Am J Psychiatry 160:284 –289, clinical trials may not show big differ- from Abbott Laboratories, Bristol-Myers-Squibb,
2003 ences in efficacy among antipsychotic Janssen, Eli Lilly, and Repligen; and holds stock in
groups, modest differences do exist, fa- Bristol-Meyers-Squibb, Eli Lilly, Merck, and Pfizer.
11. Bushe CJ, Holt R: Prevalence of diabetes J.V. has received honoraria from AstraZeneca, Bris-
and impaired glucose tolerance in pa- voring not only clozapine but also risperi- tol-Myers-Squibb, GlaxoSmithKline, and Eli Lilly
tients with schizophrenia. Br J Psychiatry done and olanzapine (5). Although the and research support from GlaxoSmithKline.
Suppl 47:S67–S71, 2004 consensus panel report acknowledged the © 2004 by the American Diabetes Association.

DIABETES CARE, VOLUME 27, NUMBER 8, AUGUST 2004 2087

Letters

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● the evidence supports the view that there prestigious consensus panel have the
References is so much disparity in the incidence of force of law in the wider professional
1. American Diabetes Association: Consen- obesity and diabetes among first- community, not to mention the pressure
sus development conference on antipsy- generation antipsychotics (FGAs) and from consumers and their advocates who
chotic drugs and obesity and diabetes
second-generation antipsychotics (SGAs). similarly take the recommendation as a
(Consensus Statement). Diabetes Care 27:
596 – 601, 2004 Furthermore, there is no evidence, for that command, considerably increasing the
2. Marder SR, Essock SM, Miller AL, Buch- matter, of disparity between the experi- exposure of prescribers to litigation.
anan RW, Davis JM, Kane JM, Lieberman ence in schizophrenia now and during the
J, Schooler NR: The Mount Sinai Confer- preantipsychotic period, especially from MICHAEL T. ISAAC, MD
ence on the pharmacotherapy of schizo- the 1920s onwards following the discov- MARIA B. ISAAC, MD, PHD
phrenia. Schizophr Bull 28:5–16, 2002 ery of insulin. It seems likely that schizo-
3. Citrome L, Jaffe A: Relationship of atypi- From The Ladywell Unit, South London and Maud-
phrenia itself represents a significant risk sley National Health Service Trust, University Hos-
cal antipsychotics with development of factor for obesity and diabetes. pital Lewisham, Lewisham High Street, London, U.K.
diabetes mellitus. Ann Pharmacother 37: Moreover, concurrent substance abuse Address correspondence to Dr. Michael T. Isaac,
1849 –1857, 2003 (especially of tobacco and cannabis) is fre- University Hospital Lewisham, South London and
4. Rosack J: FDA to require diabetes warning Maudsley National Health Service Trust, The Lady-
on antipsychotics. Psychiatr News 38(20):
quent among this patient population (2),
well Unit, London SE13 6LH, U.K. E-mail:
1, 2003 and this, along with well-established fac- [email protected].
5. Davis JM, Chen N, Glick ID: A meta-anal- tors such as low activity, sedentary life- © 2004 by the American Diabetes Association.
ysis of the efficacy of second-generation style, smoking, and, above all, poor diet
antipsychotics. Arch Gen Psychiatry 60: (with reliance on processed and “junk” ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
553–564, 2003 food), makes arguably the most signifi- References
6. Bilder RM, Goldman RS, Volavka J, Czo- cant contribution to the development of 1. American Diabetes Association: Consen-
bor P, Hoptman M, Sheitman B, Linden- obesity, dyslipidemia, and diabetes. Rela- sus development conference on antipsy-
mayer JP, Citrome L, McEvoy J, Kunz M, tive risk due to family history or ethnicity chotic drugs and obesity (Consensus State-
Chakos M, Cooper TB, Horowitz TL, is also highly relevant. ment). Diabetes Care 27:596 – 601, 2004
Lieberman JA: Neurocognitive effects of A failure to take these major con- 2. McCreadie RG: Use of drugs, alcohol, and
clozapine, olanzapine, risperidone, and tobacco by people with schizophrenia: a
haloperidol in patients with chronic
founding factors into account vitiates many case-control study. Br J Psychiatry 181:
schizophrenia or schizoaffective disorder. of the conclusions concerning the meta- 321–325, 2002
Am J Psychiatry 159:1018 –1028, 2002 bolic effects of SGAs in recent literature. We 3. Jones P: SOHO: longitudinal evidence for
7. Czobor P, Volavka J, Sheitman B, Linden- do not think that there is currently any suf- schizophrenia outcomes. Schizophr Res 67
mayer JP, Citrome L, McEvoy J, Cooper ficient evidence to distinguish among anti- (Suppl. 1):277, 2004
TB, Chakos M, Lieberman J: Antipsycho- psychotics (FGAs or SGAs), and we would
tic-induced weight gain and therapeutic not feel comfortable telling patients that a
response: a differential association. J Clin particular drug would not add to the risk of Consensus
Psychopharmacol 22:244 –251, 2002
8. Citrome L: Efficacy should drive atypical
developing diabetes. Development
We are most anxious about the rec-
antipsychotic treatment. Br Med J 326:
ommendation that “[i]f a patient gains Conference on
283–284, 2003
⬎5% of his or her initial weight at any Antipsychotic Drugs
time during therapy, one should consider
switching the SGA” (1). Weight fluctua-
and Obesity and
Consensus tion is common, can have many causes, is Diabetes
Development usually diet related (patients may be eat-
Response to consensus statement
ing more regularly in the hospital than
Conference on when in the community or they may have
Antipsychotic Drugs
I
more appetite during the recovery phase n the February 2004 issue of Diabetes
and Obesity and of their illness), and may occur after treat- Care, the American Diabetes Associa-
ment initiation but before the full thera- tion (ADA) published a summary of
Diabetes peutic effect. For example, the preliminary their conclusions drawn from the Con-
data of the large-scale SOHO study (3) of sensus Development Conference on Anti-
Response to consensus statement over 10,000 European patients indicate psychotic Drugs and Obesity and
that ⬃48% of patients put on ⬎3 kg in Diabetes (1). Although the ADA ranked

W
e of course welcome the fact that weight over 1 year of treatment. the diabetes risk for second-generation
the issue of obesity and diabetes To follow literally the consensus ad- antipsychotics (SGAs), the U.S. Food and
in the management of psychosis vice could be potentially highly detrimen- Drug Administration’s Division of Neuro-
is taken seriously by the disciplines rep- tal to the patient’s treatment. We recognize pharmacological Drug Products (DNDP)
resented in this conference (1). However, that the cost-benefit balance is duly rec- does not believe that the evidence cur-
as clinical psychiatrists we are concerned ognized and that the recommendation rently available allows such a ranking.
that some of the matters raised may give that switching should be considered does The ADA concluded that aripiprazole
an unintentionally misleading message. not command that switching must be and ziprasidone have no effect on the risk
We are far from sure, for example, that done. But even the recommendations of a of diabetes. The ADA notes that because

2088 DIABETES CARE, VOLUME 27, NUMBER 8, AUGUST 2004

 Letters

these two drugs have not been included in ted to the U.S. Food and Drug Adminis- betes Association (ADA) and other orga-
epidemiological studies, this conclusion tration do not mention weight gain as part nizations produce consensus statements
is solely based on data from clinical trials of the presentation of SGA-associated hy- (1). As stated in our clinical practice rec-
that did not reveal the risk of treatment- perglycemia or diabetes. ommendations, “the need for a consensus
emergent hyperglycemia and diabetes. Although the DNDP agrees with the statement arises when clinicians or scien-
We must point out that the clinical trial ADA’s recommendation to monitor pa- tists desire guidance on a subject for
data have not provided strong evidence of tients treated with SGAs for evidence of which there is a relative deficiency of
a diabetes risk for any of the SGAs. It is not diabetes, we do not believe that the avail- comprehensive evidence that might oth-
clear whether this is due to the timing of able evidence allows the ranking of diabe- erwise allow for a more definitive state-
glucose measurements (random in most tes risk for these drugs at this time. We ment to be made.” Therefore, it should be
cases), the low absolute frequency for di- agree with the ADA that additional stud- noted that such statements represent the
abetes events, the short duration of many ies are needed to clarify many of the issues expert opinion of the panel based on the
of the trials, or other factors. Therefore, surrounding the diabetes-SGA risk rela- presentations they heard, the literature
the DNDP does not consider the absence tionship. In the meantime, DNDP recom- they reviewed, and the considerable dis-
of a signal in clinical trial data to rule out mends that clinicians remain vigilant in cussion among panel members while
the risk of diabetes with SGAs. monitoring all patients treated with SGAs writing the statement. If there was a rea-
Based on a review of epidemiological to assure their safe use. sonable number of randomized control
studies, the ADA concluded that there is trials on the subject, there would be no
GERARD BOEHM, MD, MPH
an increased risk of diabetes with olanza- need for a consensus statement, but rather
JUDITH A. RACOOSIN, MD, MPH
pine and clozapine and discrepant results the associations would issue an official
THOMAS P. LAUGHREN, MD
with quetiapine and risperidone. The “Position Statement” or clinical guideline.
RUSSELL KATZ, MD
ADA correctly identifies many of the lim- Thus, a consensus statement can be
itations of these epidemiological studies, From the Division of Neuropharmacological Drug viewed as an expert recommendation that
Products, Center for Drug Evaluation and Research,
including “their retrospective nature, het- U.S. Food and Drug Administration, Rockville,
often precedes more definitive recom-
erogeneity of methodology, selection or Maryland. mendations issued when sufficient addi-
ascertainment bias, and absence of appro- Address correspondence to Dr. Judith A. tional data become available. These may
priate or well-characterized control sub- Racoosin, 5600 Fishers Ln., HFD-120, Rockville, or may not be different from those in the
jects, . . . relatively short periods of study, MD 20857. E-mail: [email protected]. initial consensus statement.
© 2004 by the American Diabetes Association.
. . . failure to control for a possible treat- The letter from Holt (2) comments
ment sequence bias in ‘switchover’ stud- that he was a member of another group
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
ies, and . . . not always using clinically that reviewed the evidence surrounding
References
equivalent dosages of the medications.” this issue and reached conclusions “that
1. American Diabetes Association: Consen-
The DNDP believes that although these sus development conference on antipsy- differed in some respects” from those of
studies support an increased risk of treat- chotic drugs and obesity and diabetes the ADA consensus panel. Their recom-
ment-emergent hyperglycemia or diabe- (Consensus Statement). Diabetes Care 27: mendations are due to be published soon.
tes, compared with patients treated with 596 – 601, 2004 We are, of course, delighted that another
older antipsychotic drugs, the limitations group has deliberated seriously on this
of these studies preclude firm conclusions important clinical question and look for-
about the relative risk for diabetes among Consensus ward to seeing their conclusions and rec-
the studied SGAs. Development ommendations. We hope that much new
The ADA asserts that “weight gain Conference on data come to bear on this issue in the near
and changes in body composition may ac- term. We also support Holt and each of
count for many of the purported meta- Antipsychotic Drugs the other correspondents in pointing
bolic complications associated with SGA and Obesity and out that antipsychotic medications are es-
therapy, e.g., . . . diabetes. . . ” The ADA Diabetes sential for people with schizophrenia and
correctly points out that SGAs have differ- that their differential effectiveness is
ent weight gain liabilities. Although very important. We purposefully used the
weight gain may be a factor in explaining Response to Holt, Citrome and word “consider” throughout our state-
the increased diabetes risk for SGAs, Volevka, Isaac and Isaac, and ment to suggest that the advice we gave
DNDP is not aware of evidence proving Boehm et al. with regard to avoiding undesirable side
that the treatment-emergent diabetes risk effects should be one of many factors in
for these drugs is wholly or in part due to deciding which medication to use. Such

W
treatment-emergent weight gain. Al- e appreciate the opportunity to risk-benefit considerations are essential.
though weight gain is widely recognized comment on the letters that have Citrome and Volevka (3) comment
as a risk factor for diabetes in the general been received in response to our that “the report probably overreaches
population, the clinical trial and epidemi- recent consensus statement on antipsy- available evidence when suggesting that
ological evidence has not shown a direct chotic drugs and obesity and diabetes. clinicians should consider prescribing
link between these treatment-emergent Before addressing the specific issues one antipsychotic over another with the
side effects. A substantial proportion raised in each of these letters, I think it is aim of avoiding diabetes.” We disagree
(⬃25%) of adverse event reports submit- important to note why the American Dia- with this statement and believe that the

DIABETES CARE, VOLUME 27, NUMBER 8, AUGUST 2004 2089

Letters

risk of developing diabetes must also be out that based on our analysis, the FDA Randomized Trial
considered given the devastating effects of correctly judged that there was sufficient
this disease and its complications. The ef- evidence available on the first SGAs to be- Evaluating a
ficacy of antipsychotic therapy is an ex- come clinically available to merit a warn- Predominately Fetal
tremely important factor but not the only ing label. However, we believe that there Growth–Based
one. They also express concern that our is currently insufficient epidemiologic
report might lead the inexperienced prac- data to either implicate or exculpate the Strategy to Guide
titioner to inappropriately make changes newer SGAs in this regard. We anticipate Management of
in medications, exposing the patient to in-
creased risk from an inadequately treated
that the monitoring recommended in the
consensus conference report will, if im-
Gestational Diabetes
psychiatric illness. We would hope that plemented, accelerate the accumulation in Caucasian Women
practitioners think carefully about all the of this valuable data and improve our abil-
consequences of changing medications. We ity to care for these complex patients. Response to Schaefer-Graf et al.
believe it would also be most unfortunate if In closing, our statement attempted
they ignored the possibility that a patient to bring attention to important factors re-

T
could develop one or more of the serious lated to the care of patients with psychi- he randomized trial of Schaefer-Graf
side effects associated with these drugs. atric disorders. In reaching consensus, the et al. (1) evaluating a fetal growth–
For that reason we strongly advocate and panel provided expert opinion on the based strategy to guide the manage-
carefully described a monitoring regimen. available data, including which patients ment of gestational diabetes mellitus
Isaac and Isaac (4) point out that the might be at increased risk of developing (GDM) in Berlin requires comment. The
evidence may not support the view that adverse metabolic sequelae and specify- authors used measurement of fetal ab-
there is a difference in the incidence of ing the baseline and follow-up monitor- dominal circumference (AC) at 20 –35
obesity in diabetes between those given ing that would be appropriate. We do not weeks’ gestation taken by only three ul-
first-generation antipsychotics and pa- believe our statement provides the final trasonographers (physician investigators)
tients treated with second-generation an- word on the subject. Indeed we look for- to determine the need for insulin therapy
tipsychotics (SGAs). Our panel did not ward to all future publications on the is- (36 of 90 subjects) compared with self-
address this issue since SGAs are far more sue, which will undoubtedly enhance our monitored fasting blood glucose (⬎90
widely used and preferred (with specific knowledge and provide even better clini- mg/dl) and 2-h postprandial blood glu-
exceptions noted). They also point out cal guidance. cose (⬎120 mg/dl) values obtained 2 days
that weight gain is common and can have EUGENE J. BARRETT, MD, PHD per week, which indicated the need for
many causes. We acknowledged that insulin treatment (27 of 97 subjects). The
point, but it does not obviate the concern From the Diabetes Center/Endocrinology, Univer- authors found that the ultrasound-based
that many patients on SGAs gained sub- sity of Virginia, Charlottesville, Virginia. strategy provided outcomes (12% large
stantial weight within weeks of drug ini- Address correspondence to Dr. Eugene J. Barrett, for gestational age, 17% neonatal hypo-
University of Virginia, Diabetes Center/Endocrinol-
tiation. There appears to be a differential ogy, 420 Ray C. Hunt Dr., Room 2308, Charlottes-
glycemia, 14% transfer to the neonatal in-
propensity for weight gain depending on ville, VA 22903. E-mail: [email protected]. tensive care unit [NICU]) that were no
the drug selected. © 2004 by the American Diabetes Association. different from the glycemic criteria strat-
The letter from Boehm et al. (5) sup- egy. However, the study design set up the
ports the preceding sentence, and we ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● glycemic control group for only partially
agree that proof is lacking, despite weight References intensified management by not requiring
gain being widely recognized as a risk fac- 1. American Diabetes Association: Consen- daily self-monitoring of fasting blood
tor for diabetes and our conclusion that sus development conference on antipsy- glucose (SMBG), which many other in-
the weight gain from SGAs correlates with chotic drugs and obesity and diabetes vestigators have found to produce lower
the new-onset cases of diabetes. We ap- (Consensus Statement). Diabetes Care 27: perinatal morbidity rates than in this
596 – 601, 2004
preciate that more data may be required study (2– 6), treating GDM subjects with
2. Holt RIG: Consensus development con-
before the U.S. Food and Drug Adminis- ference on antipsychotic drugs and obe- higher glucose diagnostic criteria than in
tration’s (FDA) Division of Neurophar- sity and diabetes (Letter). Diabetes Care Germany. The German authors also used
macological Drug Products itself adopts a 27:2086 –2087, 2004 tighter blood glucose treatment targets in
ranking of diabetes risk for the various 3. Citrome L, Volavka J: Consensus develop- the ultrasound-based than in the SMBG-
SGAs. However, the panel’s review of the ment conference on antipsychotic drugs based strategy.
available information, which also in- and obesity and diabetes (Letter). Diabetes The authors claimed that any physi-
cluded data presented by the FDA sup- Care 27:2087–2088, 2004 cian certified for obstetrical ultrasound
porting a differential risk of diabetes 4. Isaac MT, Isaac MB: Consensus develop- may be expected to produce reliable fetal
among the SGAs, was sufficient to unan- ment conference on antipsychotic drugs AC measurements with standard equip-
and obesity and diabetes (Letter). Diabetes
imously reach the judgment we made. ment, since their interobserver coefficient
Care 27:2088, 2004
Again, a consensus statement denotes ex- 5. Boehm G, Racoosin JA, Laughren TP, Katz of variance for the AC measurements was
pert judgement and opinion and provides R: Consensus development conference on ⬍7%. However, in most clinical settings
clinical guidance. It is understandable that antipsychotic drugs and obesity and dia- in North America, technicians make the
in order for drug labeling to change, a high- betes (Letter). Diabetes Care 27:2088 – AC measurements and physicians trained
er level of evidence is required. We point 2089. 2004 in imaging read the films. As a result,

2090 DIABETES CARE, VOLUME 27, NUMBER 8, AUGUST 2004