Psychopharmacology
DOI 10.1007/s00213-016-4411-x
REVIEW
Diabetic ketoacidosis in patients exposed to antipsychotics:
a systematic literature review and analysis of Danish
adverse drug event reports
Christoffer Polcwiartek 1,2 & Torkel Vang 1 & Christina Hedegård Bruhn 3 &
Nasseh Hashemi 1,2 & Mary Rosenzweig 3 & Jimmi Nielsen 1,2
Received: 21 March 2016 / Accepted: 11 August 2016
# Springer-Verlag Berlin Heidelberg 2016
Abstract
Rationale Patients exposed to second-generation antipsychotics
(SGAs) have approximately 10 times increased risk of diabetic
ketoacidosis (DKA) compared with the general population.
However, as DKA is a rare complication of type 2 diabetes
mellitus, and susceptible patients exposed to antipsychotics
may rapidly develop DKA independently of treatment duration
and weight gain, this is rather suggestive of type 1 diabetes
mellitus (T1DM) or latent autoimmune diabetes in adults.
Objectives We performed a systematic review of current stud-
ies regarding antipsychotic-associated DKA with type 1 etiol-
ogy and analyzed Danish adverse drug event (ADE) reports
(previously unpublished cases).
Methods PubMed, Embase, and the Cochrane Library were
searched for all relevant studies, and the Danish Medicines
Agency retrieved ADE reports using the Danish ADE data-
base (up to date as of June 28, 2016). Diagnosis of
antipsychotic-associated DKA with type 1 etiology was either
considered confirmed or possible depending on authors’ con-
clusions in the studies and/or clinical aspects. In addition,
clinico-demographic risk factors were extracted.
Results A total of 655 records and 11 ADE reports were iden-
tified, and after screening for eligibility, we included 21 case
reports/series and two ADE reports (n = 24). No relevant
clinical studies were included. Although fatal cases were
* Jimmi Nielsen
jin@rn.dk
1
Department of Psychiatry, Aalborg University Hospital,
Aalborg, Denmark
2
Department of Clinical Medicine, Aalborg University,
Aalborg, Denmark
3
The Danish Medicines Agency, Copenhagen, Denmark
identified, these were excluded because of diagnostic uncer-
tainties (n = 15). DKA occurred in 15 males (62.5 %) and nine
females (37.5 %), with a mean age ± standard deviation of
34.8 ± 12.4 years. Median time to DKA was 5 months (inter-
quartile range: 1.4–11 months). Associated antipsychotics
were olanzapine (n = 9, 36 %), aripiprazole (n = 6, 24 %),
risperidone (n = 6, 24 %), clozapine (n = 3, 12 %), and
quetiapine (n = 1, 4 %). Nine patients (37.5 %) were confirm-
edly diagnosed with T1DM following DKA resolution,
whereas 15 patients (62.5 %) had possible T1DM. In 22 pa-
tients (91.7 %), ongoing insulin treatment was required for
glycemic control.
Conclusions Increased awareness of the potential risk of
antipsychotic-associated DKA and subsequent T1DM diag-
nosis, with insulin requirements for glycemic control, is war-
ranted. The underlying mechanisms are poorly understood but
most probably multifactorial. Certainly, further studies are
warranted. Clinicians must utilize appropriate monitoring in
susceptible patients and consider the possibility of continuing
antipsychotic treatment with appropriate diabetic care.
Keywords Adverse drug event . Antipsychotics . Diabetic
ketoacidosis . Latent autoimmune diabetes in adults . Type 1
diabetes mellitus
Introduction
Diabetes mellitus and diabetic ketoacidosis associated
with antipsychotics
Both first- and second-generation antipsychotics (FGAs,
SGAs) may cause various metabolic complications, although
risk of new-onset type 2 diabetes mellitus (T2DM) is particu-
larly increased with SGA treatment (Cohen and Correll 2009).

Antipsychotic-induced weight gain is considered as an indi-
rect pathway to development of T2DM via progressive pe-
ripheral insulin resistance and decrease in insulin secretion
(Nielsen et al. 2010). While treatment with antipsychotics re-
mains a substantial risk factor for T2DM, patients with schizo-
phrenia have reduced risk of type 1 diabetes mellitus (T1DM)
(Juvonen et al. 2007). This contrast is somewhat interesting
considering that the incidence of diabetic ketoacidosis (DKA),
which occurs predominantly in connection with T1DM, is ap-
proximately 10 times higher in patients with schizophrenia ex-
posed to SGAs compared with the general population (Cohen
and Correll 2009). Some of these patients may even have undi-
agnosed diabetes mellitus (DM) reflected by increased hemoglo-
bin A1c levels for at least several weeks prior to DKA
(Henderson et al. 2007). In addition, patients with T1DM ex-
posed to antipsychotics are found to have poor glycemic control
and increased risk of severe hypoglycemia and DKA (Galler
et al. 2015).
DKA is characterized by insulinopenia, hyperglycemia,
metabolic acidosis, ketonemia, and dehydration and has been
associated with an overall mortality rate of 26.5 % in patients
exposed to antipsychotics (Cohen et al. 2005; Umpierrez and
Kitabchi 2003). It may develop at times as the first presenta-
tion of DM or, as noted, even be a late sign of long-standing
DM that was unrecognized and untreated (Nielsen et al.
2013). The underlying mechanisms of the antipsychotic-
associated DKA variant are poorly understood but most prob-
ably multifactorial. Generally, it is considered to be of type 2
etiology, where patients develop T2DM and may require oral
antidiabetic treatment (e.g., metformin) following DKA reso-
lution (Cohen 2004; Guenette et al. 2013; Jin et al. 2002,
2004). However, DKA is a rather rare complication of
T2DM unless it occurs in the setting of a trigger, including
pancreatitis, infection, myocardial infarction, or surgery
(Umpierrez and Kitabchi 2003). Adding to the complexities,
weight gain alone cannot be used as the only proxy for con-
cern regarding glucose abnormalities in patients exposed to
antipsychotics supporting the hypothesis of a more direct
antipsychotic-mediated effect on glucose metabolism or reg-
ulation (Cohen 2004; Cohen et al. 2005; Guenette et al. 2013;
Jin et al. 2002, 2004).
Overall, these features may be suggestive of type 1 etiolo-
gy, where patients initially present with DKA and subsequent-
ly develop T1DM or even latent autoimmune diabetes in
adults (LADA), with some requiring ongoing insulin treat-
ment (Cohen et al. 2005). LADA is defined by adult onset,
presence of T1DM-associated autoantibodies, and no require-
ment of insulin treatment for a period after diagnosis
(Laugesen et al. 2015). Given the pathophysiology of DKA,
particularly insufficient insulin levels to meet the body’s basic
metabolic requirements, there should be special focus on β
cell mass and function. Both FGAs and SGAs have been
associated with toxic effects on β cells rather than inhibitory
Psychopharmacology
and reversible effects as normally expected with the traditional
T2DM model (Bowen et al. 2012; Buchholz et al. 2008). This
antipsychotic-mediated toxicity could be direct and deleteri-
ous effects on β cells as well as β cell destruction triggered by
the patient’s immune system (i.e., an autoimmune process as
seen in T1DM or LADA). Such a break in tolerance is because
of abnormalities in the adaptive immune system and is known
to predate onset of T1DM or LADA.
Objectives
This study aims to identify whether antipsychotic-associated
DKA may have type 1 etiology and, in such case, identify
clinico-demographic risk factors.
Methods
Search strategy
We performed a systematic search of PubMed, Embase, and
the Cochrane Library without any time limit using the search
terms Bantipsychotic*,^ Bneuroleptic*,^ and specific names of
SGAs (i.e., amisulpride, aripiprazole, asenapine, clozapine,
iloperidone, lurasidone, olanzapine, paliperidone, quetiapine,
risperidone, sertindole, and ziprasidone) in combination with
Bketoacidosis.^ Language was restricted to English,
Scandinavian, Polish, and German. All records were imported
to the reference management software EndNote® X7.5
(Thomson Reuters, New York, NY, USA), and duplicates
were removed. Afterwards, title and abstract of all records
were sorted for relevance. In addition to the literature search,
the Danish Medicines Agency provided ADE reports that
were retrieved from the Danish ADE database (previously
unpublished cases). Search criteria were specific names of
FGAs and SGAs, Bacidosis,^ Bdiabetic ketoacidosis,^
Bdiabetic ketoacidotic hyperglycemic coma,^ Bketoacidosis,^
Bketosis,^ Blactic acidosis,^ and Bmetabolic acidosis.^ Both
searches were up to date as of June 28, 2016.
Study selection
To maintain consistency, one reviewer (first author) manually
sorted all hits based on title and abstract. Full-text studies
considered relevant for inclusion were retrieved, and two re-
viewers (first and last author) evaluated each study indepen-
dently to minimize bias. Differences in assessment were
discussed and discrepancies solved by consensus with a third
reviewer (author T. Vang). Reference lists of all included stud-
ies were also searched for additional studies. We only included
studies with relevant data on confirmed or possible
antipsychotic-associated DKA with type 1 etiology not meet-
ing any of the following exclusion criteria:
Psychopharmacology
1. Studies not regarding antipsychotic-associated DKA with
type 1 etiology
2. Data on DKA or DM type not reported
3. Death because of DKA without established DM type
4. Positive personal history of glucose intolerance/
hyperglycemia or DM prior to antipsychotic treatment
5. Not a case report/series or any kind of a prospective or
retrospective clinical study
6. Another language than specified
Data extraction
Considering that etiologies of the antipsychotic-associated
DKA variant are complex and may not always be the primary
outcome, different clinical aspects were used to determine
diagnosis of T1DM following DKA resolution. A confirmed
diagnosis depended on authors’ conclusions in the studies
and/or a positive autoantibody profile, whereas a possible di-
agnosis depended on whether the patient was discharged on
insulin and/or required insulin at follow-up. At least five types
of autoantibodies are associated with T1DM such as the clas-
sical glutamic acid decarboxylase autoantibodies, insulin au-
toantibodies, islet cell autoantibodies, tyrosine phosphatase-
related islet antigen 2 autoantibodies, and zinc transporter 8
autoantibodies (Knip and Simell 2012).
For each case of antipsychotic-associated DKA with type 1
etiology, data extraction included sex, age, ethnicity, psychiat-
ric diagnosis, antipsychotic treatment, drug-naïve status,
polypharmacy, overweight (i.e., body mass index ≥25), weight
gain during antipsychotic treatment, and family history of DM.
Specific data points regarding DKA and T1DM were also ex-
tracted, including time to onset, possible trigger, whether anti-
psychotic treatment was continued, confirmed or possible
T1DM, and T1DM treatment. Where relevant data were miss-
ing, corresponding authors of the studies were personally
contacted by email. Data are expressed as mean ± standard devi-
ation, median with interquartile range (IQR), or frequency.
Results
Overall search
A total of 655 records were identified in the literature data-
bases, and 11 ADE reports were retrieved from the Danish
ADE database (Fig. 1). After deduplication, titles and ab-
stracts of 522 records and nine ADE reports were screened.
Although we excluded 405 records, one additional record was
identified via reference list screening. Afterwards, 118 full-
text articles and nine ADE reports were screened for eligibility
yielding 21 case reports/series with data on 22 patients and
two ADE reports (n = 24) (Table 1). In addition, we identified
fatalities from DKA in both case reports/series and clinical
studies (n = 15), although these were excluded because of
several diagnostic uncertainties such as lack of established
DM diagnosis. Unfortunately, no relevant retrospective or
prospective clinical studies were identified.
Demographic and clinical background
Antipsychotic-associated DKA with type 1 etiology occurred
in 15 males (62.5 %) and nine females (37.5 %). Mean age at
admission was 34.8 ± 12.4 years (range: 6–55 years), with
four patients (16.7 %) being under 18 years of age. Where
ethnicity was reported (n = 21), nine patients (42.9 %) were
African Americans/Caribbeans and Caucasians, respectively.
Diagnostically, 12 patients (50 %) had schizophrenia or
schizoaffective disorder.
Most patients had responded inadequately to prior trials of
other antipsychotics and were also prescribed non-antipsychotic
polypharmacy at admission. Although two patients (8.3 %) were
exposed to more than one SGA concomitantly, causality was
only determined by the authors in one of the patients (olanzapine
10 mg/day, risperidone 8 mg/day). Where data on weight param-
eters were reported (n = 17), seven patients (41.2 %) manifested
no weight gain or even weight loss. However, where baseline
data on weight were reported (n = 21), 13 patients (61.9 %) were
over their ideal weight. Finally, a positive family history of DM
was only found in the minority of patients.
Antipsychotic treatment
Median time to DKA was 5 months (IQR: 1.4–11 months)
ranging from 16 days–9 years, with eight patients (33.3 %)
presenting with DKA within 2 months. Most cases were asso-
ciated with olanzapine (n = 9, 36 %), risperidone (n = 6,
24 %), and aripiprazole (n = 6, 24 %) (Table 2). On a group
level, DKA occurred more rapidly with clozapine (median: 1
month, IQR: 0.53–1.25 months) compared with olanzapine
(median: 6 months, IQR: 3.5–18 months). Antipsychotic
doses did not exceed recommended licensed ranges, with the
exception of olanzapine in one patient (>20 mg/day), and no
intentional overdose with antipsychotics or any other drugs
were reported. Finally, most patients were switched to other
antipsychotics following DKA resolution.
Outcome
Given the seriousness of DKA, immediate medical attention is
required. This includes intravenous treatment with insulin and
fluids, correction of electrolyte and metabolic derangements,
and frequent patient monitoring (Umpierrez and Kitabchi
2003). In nine patients (37.5 %), a confirmed T1DM diagnosis
was established following DKA resolution, where risperidone
most often was deemed the causative agent (n = 5) (Table 2).
 Psychopharmacology

Fig. 1 Search flow of the

Literature databases
literature databases and Danish 655 records: Danish ADE database
ADE database describing steps of • PubMed: 140 Danish Medicines Agency:
identification, screening, • Embase: 498 • 11 ADE reports
eligibility, and inclusion. Both • Cochrane Library: 17
searches were up to date as of
June 28, 2016. aAs two ADE
reports already have been 135 duplicates
excludeda
published as case reports in the
literature, they were deduplicated. 522 titles/abstracts +
b
Includes patients with a personal 9 ADE reports for screening
history of T2DM prior to DKA
(n = 9), antipsychotic-associated Total excluded: 405
DKA with type 2 etiology • Another language (26)
• Irrelevant (379)
(n = 28), transient hyperglycemia
during DKA (n = 24), and no 1 additional record passed
DKA at all (n = 19). ADE adverse reference list screening
drug event, DKA diabetic 118 full-texts +
ketoacidosis, T2DM type 2 dia- 9 ADE reports for screening
betes mellitus
Total excluded: 101 studies + 7 ADE reports
• 83 case reports/series + 7 ADE reports (136 cases):
- Underreporting (43 cases)
- Death due to DKA (13 cases)
- Not DKA type 1b (80 cases)
• 18 clinical studies:
- Underreporting (15 clinical studies)
- Death due to DKA (2 clinical studies)
- Not DKA type 1 (1 clinical study)

Passed screening for eligibility:

• Case reports: 21 (22 cases)
• ADE reports: 2
• Clinical studies: 0

The remaining 15 patients (62.5 %) were considered to have
possible T1DM, with seven patients (46.7 %) still requiring
insulin at follow-up (mean: 4.9 ± 3.4 months, range:
2 months–1 year). Finally, in five patients (20.8 %), acute
psychotic exacerbation or infection was a possible trigger of
DKA, whereas for the remaining 19 patients (79.2 %), no
apparent triggers of DKA were identified apart from the pos-
sible association with antipsychotics.
Discussion
Implications for clinical care
This study was performed to address the lack of studies in the
literature characterizing patients with antipsychotic-associated
DKA with type 1 etiology. To better understand the involved
clinico-demographic risk factors, we reviewed cases both pub-
lished in the literature and registered in the Danish ADE da-
tabase (previously unpublished cases). Our key finding is that
more than one third of the total reviewed cases were subse-
quently diagnosed with T1DM following DKA resolution,
whereas the remaining cases were considered as possible
T1DM. For these cases, type 1 etiology could not definitely
be ruled out, as they were discharged on insulin, and
approximately half of them still required insulin at follow-
up. However, this can also reflect a severe T2DM that did
not respond to oral antidiabetics, or the prescriber’s perception
that patients required ongoing insulin treatment. Overall, this
raises the question of whether antipsychotics may be associ-
ated with an autoimmune β cell destruction, thereby causing
T1DM, where susceptible patients may initially present with
DKA.
Our findings should be interpreted within their limitations.
Importantly, no retrospective or prospective clinical studies were
identified, and cases were most likely subject to publication bias
because only the most extraordinary may have reached threshold
of publication. In addition to underreporting, the retrospective
nature of case data also leaves certain questions unanswered such
as pre-DM prior to DKA. Some patients could also have devel-
oped DKA with type 1 etiology regardless of exposure to anti-
psychotics. This may especially be true for the younger patients,
where the overall risk of developing T1DM is at the highest. In
addition, more than half of the included patients were prescribed
polypharmacy at time of DKA making it difficult to determine
actual causality of any particular antipsychotic.
For clinicians, early identification of DKA is crucial to
ensure that appropriate treatment is initiated as quickly as
possible. It appears that most SGAs are at potential risk of
causing DKA with type 1 etiology at any time following
Table 1 Summary of clinico-demographic risk factors of patients with antipsychotic-associated DKA with type 1 etiology (n = 24)

Reference Case Treatment status Risk factors DKA and T1DM

Psychopharmacology

Sex/age, Ethnicity/ SGA, mg/day Drug Polypharmacy Overweight Weight DM in Onset, Trigger SGA Diagnosisa Treatment
years diagnosis naïve gain family month(s) treatment

(Peterson and Byrd M/46 AA/SCZ CLO 500 No Yes N/R No Yes (type N/R) 1.25 No Discont’d Conf. Ins.
1996)
(Goldstein et al. 1999) F/42 CC/SZA OLA 10 No N/R No Yes T2DM 6 No Switched Poss. Ins.
(Smith et al. 1999) M/40 AC/SCZ CLO (dose N/R) No N/R Yes N/R No 0.53 Exa. N/R Poss. Ins.
(Croarkin et al. 2000) M/42 CC/DEPR RIS 4 No Yes No No No 6 No Switched Conf. Ins.
(Dahri and Brown M/22 CC/SCZ OLA 10 and RIS Yes Yes Yes N/R Yes (type N/R) 2 No Switched Poss. Ins.
2002) 8
(Tavakoli and Arguisola M/35 CC/BD OLA 5 No Yes Yes Yes Yes (type N/R) 18 No Discont’d Conf. Ins.
2003)
(Wilson et al. 2003) M/33 AA/SZA CLO 550 and No Yes Yes Yes No 1 No Cont’d Poss. Ins.
OLA 10
M/48 AA/SCZ OLA 30 No No Yes Yes No 10 No Cont’d Poss. Ins.
(Howes and Rifkin F/41 AC/SZA OLA 20 No Yes Yes Yes T2DM 3.5 Inf. Cont’d Poss. Ins.
2004)
(Babu et al. 2005) F/15 N/R, BD ARI (dose N/R) N/R Yes N/R N/R N/R 4 No Discont’d Poss. Ins.
(Bahceci et al. 2005) M/37 CC/BD RIS 4 Yes Yes Yes Yes No 6 No Cont’d Conf. No
(Reddymasu et al. M/33 AA/SCZ ARI (dose N/R) N/R N/R Yes Yes No 18 No Switched Poss. Ins.
2006)
(Varma et al. 2007) F/35 CC/BD OLA 10 Yes No Yes Yes No 1.5 Exa. and inf. Switched Poss. Ins.
(Wong et al. 2007) M/22 Asian/SCZ OLA 10 No Yes No Yes No 39 No Switched Poss. DD and ins.
(Makhzoumi et al. M/44 AA/SZA ARI 30 No Yes Yes No No 0.53 Exa. Switched Poss. OA and ins.
2008)
(Sato et al. 2008) F/46 Asian/SCZ RIS 3 No No No No Yes (type N/R) 4 No Switched Conf. Ins.
(Sirois 2008) F/41 AA/DEPR QUE 400 No Yes Yes N/R No 1.1 No N/R Poss. Ins.
(Nwosu et al. 2009) F/15 CC/BD ARI 15 Yes Yes No No T1DM 12 No Discont’d Conf. Ins.
(Nahas et al. 2010) F/45 N/R, BD OLA (dose N/R) N/R N/R N/R N/R Yes (type N/R) 36 No Switched Poss. Ins.
(Sæverud and Gerlyng M/42 CC/DEPR OLA 7.5 Yes Yes Yes N/R N/R 6 Inf. Cont’d Conf. No
2010)
ADE report #1 (2011) F/6 Asian/ADHD RIS 0.5 Yes Yes No No GDM 1.1 No Cont’d Conf. Ins.
(Patel et al. 2011) M/32 N/R, SCZ ARI (dose N/R) No N/R Yes N/R N/R 3 No Discont’d Poss. OA and ins.
(Watkins et al. 2011) M/55 AA/DEPR ARI 10 Yes Yes No Yes N/R 6 No Discont’d Poss. DD and ins.
ADE report #2 (2015) M/17 CC/TS RIS 0.5 Yes No No No No 108 No Discont’d Conf. Ins.

a
A confirmed T1DM diagnosis following DKA resolution depended on authors’ conclusions in the studies and/or a positive autoantibody profile, whereas a possible diagnosis depended on whether the
patient was discharged on insulin and/or required insulin at follow-up
AA African American, AC African Caribbean, ADE adverse drug event, ARI aripiprazole, BD bipolar disorder, CC Caucasian, CLO clozapine, Conf. confirmed, DD diabetic diet, DEPR depression, DKA
diabetic ketoacidosis, DM diabetes mellitus, Exa. acute psychotic exacerbation, F female, GDM gestational diabetes mellitus, Inf. infection, Ins. insulin, M male, N/R not reported, OA oral antidiabetic, OLA
olanzapine, Poss. possible, QUE quetiapine, RIS risperidone, SCZ schizophrenia, SGA second-generation antipsychotic, SZA schizoaffective disorder, T1DM type 1 diabetes mellitus, T2DM type 2 diabetes
mellitus, TS Tourette syndrome
 Psychopharmacology

Table 2 Overall comparison of

antipsychotic-associated DKA SGA DKA, Confirmed SCZ or M:F Mean Median time to Dose
with type 1 etiology, sorted by n (%) T1DM, n SZA, n ratio age ± SD DKA (IQR), range, mg/
SGA (n = 25)a (%) (%) (range), years months day (n)

Aripiprazole 6 (24) 1 (16.7) 3 (50) 4:2 32.3 ± 15.8 5 (3–12) 10–30 (3)
(15–55)
Clozapine 3 (12) 1 (33.3) 3 (100) 3:0 39.7 ± 6.5 1 (0.53–1.25) 500–550
(33–46) (2)
Olanzapine 9 (36) 2 (22.2) 5 (55.6) 5:4 36.9 ± 9.4 6 (3.5–18) 5–30 (8)
(22–48)
Quetiapine 1 (4) 0 0 0:1 41 1.1 400
Risperidone 6 (24) 5 (83.3) 3 (50) 4:2 30.2 ± 15.5 5 (1.1–6) 0.5–8
(6–46)
a
Although two patients were exposed to more than one SGA concomitantly, causality was only determined by
the authors in one of the patients (olanzapine 10 mg/day, risperidone 8 mg/day)
DKA diabetic ketoacidosis, IQR interquartile range, M:F male:female, SCZ schizophrenia, SD standard deviation,
SGA second-generation antipsychotic, SZA schizoaffective disorder, T1DM type 1 diabetes mellitus

initiation of antipsychotic treatment, with the exception of,
e.g., ziprasidone that only has been associated with severe
hyperglycemia (Guenette et al. 2013). With the exception of
olanzapine, most cases were associated with the low-risk met-
abolic SGAs aripiprazole and risperidone, instead of SGAs
with more potent diabetogenic effects such as clozapine that
most frequently has been associated with DM (Nielsen et al.
2013). Although incidence of DKA with clozapine is general-
ly low (0.1–0.3 %), risk of mortality varies from 20–31 %
(Cohen et al. 2012). Particularly, in other reviews of
antipsychotic-associated DKA, clozapine has been associated
with most cases, but etiologies of DKA in these reviews were
not clearly distinguished (i.e., transient hyperglycemia during
DKA, T1DM, or T2DM) (Cohen 2004; Guenette et al. 2013;
Jin et al. 2002). As aripiprazole is less likely to be associated
with weight gain and glucose abnormalities, the associated
cases are rather suggestive of type 1 etiology (Nielsen et al.
2010). However, a confounding by indication cannot be ruled
out because patients at high risk for metabolic complications
may more likely have aripiprazole prescribed because of safe-
ty reasons (Nielsen et al. 2010; Polcwiartek et al. 2015).
Finally, clinicians should also be aware that rare DKA cases
have been associated with FGAs and other psychotropics such
as lithium (Guenette et al. 2013).
Most age groups of patients exposed to antipsychotics are
at potential risk of developing DKA with type 1 etiology. In
comparison, majority of patients with T1DM in the general
population are diagnosed before 30 years of age, with most
patients being children and adolescents. A potential explana-
tion for this discrepancy may be that antipsychotics are rarely
prescribed before 15 years of age. In addition, T1DM can
become clinically apparent at any age. For example, twin studies
have shown that discordant identical twins of patients with
T1DM are at high risk of developing T1DM-associated autoan-
tibodies and subsequently progress to overt T1DM. The lag
phase in this regard may be as long as 30 years indicating im-
portant environmental contributions (van Belle et al. 2011). It is
also well known that some patients >30 years of age slowly can
develop T1DM, a variant of T1DM referred to as LADA
(Laugesen et al. 2015).
Regarding T1DM in the general population, the
male:female ratio is approximately 1.5:1 (Gale and Gillespie
2001). While DKA has been reported to occur twice as fre-
quently in females in the general population (Umpierrez and
Kitabchi 2003), this finding has later been contradicted by an
extensive meta-analysis showing no sex difference (Usher-
Smith et al. 2011). However, we found that most cases oc-
curred in males, which corresponds to findings in the reviews
without distinguished etiologies (Cohen 2004; Guenette et al.
2013; Jin et al. 2002). Along similar lines, these reviews also
found that most cases occurred in African Americans/
Caribbeans indicating an overrepresentation of these ethnic
groups compared with larger ethnic groups such as
Caucasians. This corresponds with previous studies of DKA in
the general population, where there seems to be an overrepresen-
tation of patients from minority groups (Usher-Smith et al. 2011).
As noted, the overall mortality rate of DKA in patients
exposed to antipsychotics is 26.5 % compared with the mor-
tality rate of <5 % that has been reported in the general pop-
ulation (Cohen et al. 2005; Umpierrez and Kitabchi 2003). In
addition, we identified 15 fatalities from DKA that were ex-
cluded because of several diagnostic uncertainties, as well as
7.3 % of cases were fatal in one of the reviews without distin-
guished etiologies (Guenette et al. 2013). This underscores the
life-threatening potential of DKA in patients exposed to anti-
psychotics. However, because of its rather silent autopsy na-
ture, fatal DKA may be neglected as a potential cause of
sudden death, and risk of mortality may even be higher than
reported here (Ali et al. 2012). Usually, sudden death in pa-
tients exposed to antipsychotics is multifactorial and may as
Psychopharmacology
well be caused by cardiac complications, including torsades
de pointes, myocardial infarction, myocarditis, and tardive
dyskinesia-induced arrhythmia (Polcwiartek et al. 2016).
Focus on autoimmune mechanisms
Genetic factors play a major role in the autoimmune nature of
T1DM, with an estimated heritability of ≤88 % and a concor-
dance rate of monozygotic twins of approximately 50 %
(Hyttinen et al. 2003; Ikegami and Ogihara 1996). In general,
studies indicate that patients with schizophrenia are less prone
to develop T1DM and probably other autoimmune diseases
such as rheumatoid arthritis (Juvonen et al. 2007). Such auto-
immune diseases are characterized by exaggerated T helper 1
(Th1) responses and associated cytokines such as interferon-γ
(IFN-γ) and tumor necrosis factor-α (TNF-α). It has been
demonstrated that patients with schizophrenia display a
skewing in their T helper responses, where Th1 responses
generally are weak, and T helper 2 (Th2) responses are rela-
tively stronger (Riedel et al. 2007). This provides an explana-
tion for the lower risk of T1DM in patients with schizophre-
nia. In addition, prevalence of T1DM-associated autoanti-
bodies in patients with schizophrenia is similar to that in the
general population and even patients with T2DM indicating
that DM in patients with schizophrenia is less likely to be
autoimmune mediated (i.e., T1DM or LADA) (Cohen et al.
2005).
As noted, the incidence of DKA is approximately 10 times
higher in patients exposed to SGAs, where most have a psy-
chotic disorder, compared with the general population (Cohen
and Correll 2009). The latter finding is rather surprising con-
sidering that DKA most often develops in connection with
T1DM. In our perspective, there are three potential mecha-
nisms that can explain antipsychotic-associated DKA with
type 1 etiology, but they are not mutually exclusive:
1. Neglect of early symptoms of T1DM in patients exposed
to antipsychotics
2. Activation of the immune system associated with acute
psychotic exacerbation that is treated with antipsychotics
3. Immunomodulatory effects of antipsychotics that may
cause loss of tolerance and autoimmune β cell destruction
Compared with the general population, early symptoms of
T1DM may more often be neglected in patients exposed to
antipsychotics. This is probably because most patients that are
being prescribed antipsychotics suffer from a psychotic disor-
der. As a consequence, physicians may tend to overlook the
diagnosis of T1DM in these patients and, because of the psy-
chotic disorder, these patients are more likely to ignore or
misinterpret their symptoms of T1DM. In addition, level of
integration in society and educational level for these patients
may also play a role in the diagnostic delay. In agreement with
this, studies of DKA in the general population have suggested
an overrepresentation of patients from minority groups
(Usher-Smith et al. 2011). Thereby, the delay in the diagnosis
of T1DM for the reasons described above may contribute
substantially to the increased incidence of DKA in patients
exposed to SGAs compared with the general population.
Here, however, there are several potential confounding factors
that also need to be discussed.
During acute exacerbation of psychotic disorders such as
schizophrenia, there is a systemic response with activation of
the immune system and concomitantly elevated plasma levels
of several cytokines. Some of these cytokines can serve as
state markers such as interleukin (IL)-1β, IL-6, and
transforming growth factor β, as they normalize in response
to treatment with antipsychotics. Other cytokines are unaffect-
ed by treatment and therefore are referred to as trait markers,
including IFN-γ and TNF-α (Røge et al. 2012). Considering
the remitting-relapsing nature of the autoimmune β cell de-
struction by T cells prior to development of clinically relevant
T1DM, a state of hypercytokinemia in susceptible patients could
potentially tip the balance resulting in rapid progression towards
T1DM. In addition to such bystander activation of autoreactive T
cells, exposure of β cells to certain cytokines, including IFN-γ,
IL-1β, and TNF-α, may render them more vulnerable for both
apoptosis and autoimmune destruction, overall causing clinically
significant reduction of β cell mass (Lopes et al. 2014; Meier
et al. 2006; Seewaldt et al. 2000; van Belle et al. 2011).
Antipsychotics interact with various cell surface receptors, of
which dopamine receptors (DRs) are the most important ones.
DRs are heptahelical transmembrane proteins that generate intra-
cellular signals mainly via interaction with G proteins. There are
five different DRs and they are grouped into two families, D1-
like (encompassing D1 and D5) and D2-like (encompassing D2–4)
(Neve et al. 2004). D1-like receptors are coupled to the stimula-
tory Gα subunit, which via activation of adenylyl cyclase leads to
elevated cAMP levels. D1-like receptors may also activate phos-
pholipase C. In contrast, D2-like receptors are coupled to the
inhibitory Gα subunit that may affect intracellular signaling in
several ways, most importantly via inhibition of adenylyl cy-
clase. In addition to the role of dopamine as a transmitter sub-
stance in the central nervous system, it plays a role in several
other organ systems. For example, certain cell types in the im-
mune system, including dendritic cells (DCs) and regulatory T
cells, can synthesize and secrete dopamine, which then via auto-
crine and paracrine mechanisms can exert important immunoreg-
ulatory effects. Most T cell subsets express all five subtypes of
DRs, whereas DCs express all subtypes apart from D 4
(Kustrimovic et al. 2014; Prado et al. 2012).
Dopamine is considered to affect CD4+ T helper cell dif-
ferentiation and have immunostimulatory properties (e.g., via
activation of effector T cells, with CD8+ T cells more than
CD4+ T cells, and inhibition of regulatory T cells) (Franz
et al. 2015; Levite 2016). However, overall effects of

dopamine on cells of the immune system may be difficult to
predict because DR expression levels depend on cell type and
level of cell activation/differentiation. In addition, local dopa-
mine concentration is important because affinity of DRs for
dopamine varies (D3 >D5 >D4 >D2 >D1), and different DRs
may be coupled to opposing intracellular signaling pathways
(Pacheco et al. 2014). Along these lines, immunosuppressive
effects of various DR agonists have also been reported (Huang
et al. 2010). As antipsychotics are DR antagonists (i.e., most
importantly D2-like receptor antagonists) and given the com-
plexity of dopamine-mediated effects on the immune system,
it is difficult to present an accurate model of how antipsy-
chotics affect the immune system. In addition, SGAs may also
interact with serotonin 5-HT receptors (e.g., 5-HT2A), which
together with several other serotonin 5-HT receptors are
expressed in immune cells such as CD4+ and CD8+ T cells
(Martins et al. 2012). It is therefore not surprising that anti-
psychotics have been associated with both
immunostimulatory and immunosuppressive effects. This is
at least the case for clozapine, haloperidol, and risperidone
(Baumeister et al. 2016).
Based on this, it is difficult to define the underlying
pathophysiological mechanisms of antipsychotic-
associated DKA with type 1 etiology. One intriguing mod-
el would be the following: Patients with a chronic psy-
chotic disorder originally display a skewing in their T cell
responses (i.e., lowered Th1:Th2 ratio), which explains
their reduced risk for autoimmune diseases such as
T1DM and rheumatoid arthritis (Juvonen et al. 2007;
Pacheco et al. 2014; Riedel et al. 2007). In connection
with acute psychotic exacerbation, a systemic activation
of the immune system appears, but initiation of treatment
with antipsychotics may also lead to elevated levels of IL-
12 favoring Th1 responses. This is at least the case for
drug-naïve patients experiencing a first episode of psy-
chosis (Crespo-Facorro et al. 2008).
IL-12-associated mechanisms, perhaps accompanied by
systemic immune activation caused by the acute psychotic
exacerbation per se, may generate aberrant Th1 responses
that in susceptible patients exposed to antipsychotics can
cause loss of tolerance and subsequent autoimmune β cell
destruction. The end-result will be T1DM that in situa-
tions with significant neglect may present as DKA. The
proposed model is supported by data indicating that D2
and D3 agonists, which have an opposite receptor profile
compared with antipsychotics, may induce tolerance and
be useful in the treatment of patients with autoimmune
diseases, including systemic lupus erythematosus and
rheumatoid arthritis (Pacheco et al. 2014). However, we
need to underscore that the presented model lacks firm
scientific support because there are currently too many
contradictory findings regarding the immunological ef-
fects of antipsychotics.
Psychopharmacology
Conclusions
Susceptible patients exposed to antipsychotics, especially
SGAs, may develop DKA and subsequently be diagnosed
with T1DM, where ongoing insulin treatment for glycemic
control is indicated. This may occur independently of treat-
ment duration, weight gain, and other known risk factors for
DKA. Certainly, more studies assessing the pathophysiology
of the antipsychotic-associated DKA variant are warranted. In
the meantime, clinicians must remain vigilant given its acute
onset and potential lethality and keep in mind that it is possible
to continue antipsychotic treatment with appropriate diabetic
care.
Compliance with ethical standards
Conflict of interest CP has received speaking fees from H. Lundbeck.
JN has received speaking fees from Bristol-Myers Squibb, H. Lundbeck,
and Hemocue, as well as research grants from H. Lundbeck and Pfizer for
clinical studies. All other authors declare that they have no conflict of
interest.
References
Ali Z, Levine B, Ripple M, Fowler DR (2012) Diabetic ketoacidosis: a
silent death. Am J Forensic Med Pathol 33:189–193
Babu KM, Ganetsky M, Liang IE, Bird SB, Boyer EW (2005)
Pancreatitis and diabetic ketoacidosis. Clin Toxicol (Phila) 43:642
Bahceci M, Tuzcu A, Bulent C, Cengiz T (2005) Risperidone-associated
transient diabetic ketoacidosis and diabetes mellitus type 1 in a pa-
tient treated with valproate and lithium. Pharmacopsychiatry 38:
105–106
Baumeister D, Ciufolini S, Mondelli V (2016) Effects of psychotropic
drugs on inflammation: consequence or mediator of therapeutic ef-
fects in psychiatric treatment? Psychopharmacology 233:1575–
1589
Bowen E, Holt R, Harrison M (2012) β-cell-specific effects of the anti-
psychotic drugs clozapine and haloperidol. Diabet Med 29(Suppl 1):
32
Buchholz S, Morrow AF, Coleman PL (2008) Atypical antipsychotic-
induced diabetes mellitus: an update on epidemiology and postulat-
ed mechanisms. Intern Med J 38:602–606
Cohen D (2004) Atypical antipsychotics and new onset diabetes
mellitus. An overview of the literature. Pharmacopsychiatry
37:1–11
Cohen D, Correll CU (2009) Second-generation antipsychotic-associated
diabetes mellitus and diabetic ketoacidosis: mechanisms, predictors,
and screening need. J Clin Psychiatry 70:765–766
Cohen D, Batstra MR, Gispen-de Wied CC (2005) Immunological char-
acteristics of diabetes in schizophrenia. Diabetologia 48:1941–1942
Cohen D, Bogers JP, van Dijk D, Bakker B, Schulte PF (2012) Beyond
white blood cell monitoring: screening in the initial phase of cloza-
pine therapy. J Clin Psychiatry 73:1307–1312
Crespo-Facorro B, Carrasco-Marín E, Pérez-Iglesias R, Pelayo-Terán JM,
Fernandez-Prieto L, Leyva-Cobián F, Vázquez-Barquero JL (2008)
Interleukin-12 plasma levels in drug-naïve patients with a first epi-
sode of psychosis: effects of antipsychotic drugs. Psychiatry Res
158:206–216
Psychopharmacology
Croarkin PE, Jacobs KM, Bain BK (2000) Diabetic ketoacidos is associ-
ated with risperidone treatment? Psychosomatics 41:369–370
Dahri K, Brown G (2002) Combined olanzapine- and risperidone-
induced diabetic ketoacidosis. Can J Hosp Pharm 55:282–283
Franz D, Contreras F, Gonzalez H, Prado C, Elgueta D, Figueroa C,
Pacheco R (2015) Dopamine receptors D3 and D5 regulate CD4(+
)T-cell activation and differentiation by modulating ERK activation
and cAMP production. J Neuroimmunol 284:18–29
Gale EA, Gillespie KM (2001) Diabetes and gender. Diabetologia 44:3–
15
Galler A, Bollow E, Meusers M, Bartus B, Nake A, Haberland H,
Schober E, Holl RW (2015) Comparison of glycemic and metabolic
control in youth with type 1 diabetes with and without antipsychotic
medication: analysis from the nationwide German/Austrian
Diabetes Survey (DPV). Diabetes Care 38:1051–1057
Goldstein LE, Sporn J, Brown S, Kim H, Finkelstein J, Gaffey GK, Sachs
G, Stern TA (1999) New-onset diabetes mellitus and diabetic
ketoacidosis associated with olanzapine treatment. Psychosomatics
40:438–443
Guenette MD, Hahn M, Cohn TA, Teo C, Remington GJ (2013) Atypical
a n t i p s y c h o t i c s a n d d i a b e t i c k e t o a c i d o s i s : a r e v i e w.
Psychopharmacology 226:1–12
Henderson DC, Cagliero E, Copeland PM, Louie PM, Borba CP, Fan X,
Freudenreich O, Goff DC (2007) Elevated hemoglobin A1c as a
possible indicator of diabetes mellitus and diabetic ketoacidosis in
schizophrenia patients receiving atypical antipsychotics. J Clin
Psychiatry 68:533–541
Howes OD, Rifkin L (2004) Diabetic keto-acidotic (DKA) coma follow-
ing olanzapine initiation in a previously euglycaemic woman and
successful continued therapy with olanzapine. J
Psychopharmacol 18:435–437
Huang Y, Qiu AW, Peng YP, Liu Y, Huang HW, Qiu YH (2010) Roles of
dopamine receptor subtypes in mediating modulation of T lympho-
cyte function. Neuro Endocrinol Lett 31:782–791
Hyttinen V, Kaprio J, Kinnunen L, Koskenvuo M, Tuomilehto J (2003)
Genetic liability of type 1 diabetes and the onset age among 22,650
young Finnish twin pairs: a nationwide follow-up study. Diabetes
52:1052–1055
Ikegami H, Ogihara T (1996) Genetics of insulin-dependent diabetes
mellitus. Endocr J 43:605–613
Jin H, Meyer JM, Jeste DV (2002) Phenomenology of and risk factors for
new-onset diabetes mellitus and diabetic ketoacidosis associated
with atypical antipsychotics: an analysis of 45 published cases.
Ann Clin Psychiatry 14:59–64
Jin H, Meyer JM, Jeste DV (2004) Atypical antipsychotics and
glucose dysregulation: a systematic review. Schizophr Res
71:195–212
Juvonen H, Reunanen A, Haukka J, Muhonen M, Suvisaari J, Arajärvi R,
Partonen T, Lönnqvist J (2007) Incidence of schizophrenia in a
nationwide cohort of patients with type 1 diabetes mellitus. Arch
Gen Psychiatry 64:894–899
Knip M, Simell O (2012) Environmental triggers of type 1 diabetes. Cold
Spring Harb Perspect Med 2:a007690
Kustrimovic N, Rasini E, Legnaro M, Marino F, Cosentino M (2014)
Expression of dopaminergic receptors on human CD4+ T lympho-
cytes: flow cytometric analysis of naive and memory subsets and
relevance for the neuroimmunology of neurodegenerative disease. J
NeuroImmune Pharmacol 9:302–312
Laugesen E, Østergaard JA, Leslie RD (2015) Latent autoimmune diabe-
tes of the adult: current knowledge and uncertainty. Diabet Med 32:
843–852
Levite M (2016) Dopamine and T cells: dopamine receptors and potent
effects on T cells, dopamine production in T cells, and abnormalities
in the dopaminergic system in T cells in autoimmune, neurological
and psychiatric diseases. Acta Physiol (Oxf) 216:42–89
Lopes M, Kutlu B, Miani M, Bang-Berthelsen CH, Størling J, Pociot F,
Goodman N, Hood L, Welsh N, Bontempi G, Eizirik DL (2014)
Temporal profiling of cytokine-induced genes in pancreatic beta-
cells by meta-analysis and network inference. Genomics 103:264–
275
Makhzoumi ZH, McLean LP, Lee JH, Ibe AI (2008) Diabetic
ketoacidosis associated with aripiprazole. Pharmacotherapy 28:
1198–1202
Martins LCA, Rocha NP, Torres KCL, dos Santos RR, França GS, de
Moraes EN, Mukhamedyarov MA, Zefirov AL, Rizvanov AA,
Kiyasov AP, Vieira LB, Guimarães MM, Yalvaç ME, Teixeira AL,
Bicalho MAC, Janka Z, Romano-Silva MA, Palotás A, Reis HJ
(2012) Disease-specific expression of the serotonin-receptor 5-
HT(2C) in natural killer cells in Alzheimer’s dementia. J
Neuroimmunol 251:73–79
Meier JJ, Ritzel RA, Maedler K, Gurlo T, Butler PC (2006) Increased
vulnerability of newly forming beta cells to cytokine-induced cell
death. Diabetologia 49:83–89
Nahas M, Krishna V, Lohiya S (2010) Diabetic ketoacidosis presenting in
an ‘atypical’ way. J Gen Intern Med 25(Suppl 3):S507
Neve KA, Seamans JK, Trantham-Davidson H (2004) Dopamine recep-
tor signaling. J Recept Signal Transduct Res 24:165–205
Nielsen J, Correll CU, Manu P, Kane JM (2013) Termination of clozapine
treatment due to medical reasons: when is it warranted and how can
it be avoided? J Clin Psychiatry 74:603–613 quiz 613
Nielsen J, Skadhede S, Correll CU (2010) Antipsychotics associated with
the development of type 2 diabetes in antipsychotic-naïve schizo-
phrenia patients. Neuropsychopharmacology 35:1997–2004
Nwosu BU, Hardy OT, Angelescu A, Soyka LA, Lee MM (2009) Do
atypical antipsychotic agents trigger autoimmune diabetes?
Endocrinologist 19:85–87
Pacheco R, Contreras F, Zouali M (2014) The dopaminergic system in
autoimmune diseases. Front Immunol 5:117
Patel VB, Bulchandani DG, Kyner JL (2011) Severe insulin resistance
associated with atypical antipsychotic agents. Diabetes 60(Suppl 1):
A681
Peterson GA, Byrd SL (1996) Diabetic ketoacidosis from clozapine and
lithium cotreatment. Am J Psychiatry 153:737–738
Polcwiartek C, Kragholm K, Schjerning O, Graff C, Nielsen J (2016)
Cardiovascular safety of antipsychotics: a clinical overview.
Expert Opin Drug Saf 15:679–688
Polcwiartek C, Sneider B, Graff C, Taylor D, Meyer J, Kanters JK,
Nielsen J (2015) The cardiac safety of aripiprazole treatment in
patients at high risk for torsade: a systematic review with a meta-
analytic approach. Psychopharmacology 232:3297–3308
Prado C, Contreras F, González H, Díaz P, Elgueta D, Barrientos
M, Herrada AA, Lladser Á, Bernales S, Pacheco R (2012)
Stimulation of dopamine receptor D5 expressed on dendritic
cells potentiates Th17-mediated immunity. J Immunol 188:
3062–3070
Reddymasu S, Bahta E, Levine S, Manas K, Slay LE (2006) Elevated
lipase and diabetic ketoacidosis associated with aripiprazole. JOP 7:
303–305
Riedel M, Spellmann I, Schwarz MJ, Strassnig M, Sikorski C, Möller HJ,
Müller N (2007) Decreased T cellular immune response in schizo-
phrenic patients. J Psychiatr Res 41:3–7
Røge R, Møller BK, Andersen CR, Correll CU, Nielsen J (2012)
Immunomodulatory effects of clozapine and their clinical implica-
tions: what have we learned so far? Schizophr Res 140:204–213
Sæverud HA, Gerlyng P (2010) From prison with coma. Tidsskr Nor
Legeforen 130:284–285
Sato Y, Yasui-Furukori N, Kaneko S, Moriyama T (2008) New-onset
diabetic ketoacidosis in a schizophrenic patient with multiple auto-
immune disease during treatment with risperidone. Prog
Neuropsychopharmacol Biol Psychiatry 32:577–578

Seewaldt S, Thomas HE, Ejrnaes M, Christen U, Wolfe T, Rodrigo E,
Coon B, Michelsen B, Kay TW, von Herrath MG (2000) Virus-
induced autoimmune diabetes: most beta-cells die through inflam-
matory cytokines and not perforin from autoreactive (anti-viral) cy-
totoxic T-lymphocytes. Diabetes 49:1801–1809
Sirois F (2008) New-onset diabetic ketoacidosis associated with
quetiapine: a case report. Gen Hosp Psychiatry 30:587–588
Smith H, Kenney-Herbert J, Knowles L (1999) Clozapine-induced dia-
betic ketoacidosis. Aust N Z J Psychiatry 33:120–121
Tavakoli SA, Arguisola MS (2003) Diabetic ketoacidosis in a patient
treated with olanzapine, valproic acid, and venlafaxine. South Med
J 96:729–730
Umpierrez GE, Kitabchi AE (2003) Diabetic ketoacidosis: risk factors
and management strategies. Treat Endocrinol 2:95–108
Usher-Smith JA, Thompson MJ, Sharp SJ, Walter FM (2011) Factors asso-
ciated with the presence of diabetic ketoacidosis at diagnosis of diabetes
in children and young adults: a systematic review. BMJ 343:d4092
Psychopharmacology
van Belle TL, Coppieters KT, von Herrath MG (2011) Type 1 diabetes:
etiology, immunology, and therapeutic strategies. Physiol Rev 91:
79–118
Varma MK, Connolly K, Fulton B (2007) Life-threatening hyperglyce-
mia and acidosis related to olanzapine: a case report and review of
the literature. J Intensive Care Med 22:52–55
Watkins J, Ray S, Gillion A, Finch C (2011) Case report - aripiprazole-
induced diabetic ketoacidosis: remaining vigilant. Hosp Pharm 46:
432–435
Wilson DR, D’Souza L, Sarkar N, Newton M, Hammond C (2003) New-
onset diabetes and ketoacidosis with atypical antipsychotics. Schizophr
Res 59:1–6
Wong JO, Fu JC, Hung GB (2007) Olanzapine-induced diabetic
ketoacidosis in a Chinese man. Hong Kong Med J 13:73–74