Introduction
Management of diabetes lends itself very well to several classical rules of good clinical
practice, an elaboration of which is out of the scope of this short story. However,
a couple of these are noteworthy. “Diagnosis should precede treatment as much as possible
except for measures of resuscitation” applies to diabetes as to all other conditions.
In diabetes, diagnosis and classification are the same. Hence, a serious attempt to
classify diabetes at the time of diagnosis or as soon as possible after that is mandatory.[1] In particular, when the patient's characteristics are atypical, or events do not
follow the expected course, making assumptions under these circumstances can be very
dangerous.[2] In this vignette, an unusual case of diabetes in a young woman is presented and
discussed with an analysis of lessons to be learned. The present case report exemplifies
several themes of “not expecting the expected,” “not making unfounded assumptions,”
and “ignoring several alert signals.”
In the Merriam-Webster dictionary, the idiom “in the nick of time” means just before
the last moment when something can be changed or something terrible will happen.[3] Many examples are medical, for instance, “The ambulance arrived in the nick of time”
or “The doctor arrived in the nick of time. The patient's life was saved”. Hence,
the choice of the title is not for fun, but it is perhaps the best description of
the case as the story unfolds.
Case Report
The Scene
On a very busy Thursday, the on-call endocrinologist was about to start seeing his
patient in the diabetic clinic when the clinic nurse at the door asked if he could
accommodate the 9 am patient who arrived late at 11:30 am from another town 265 km.
His cellular phone and clinic phone rang at the same time. He pretended to be in control
while trying to sort out all these issues simultaneously; he apologized to the patient
for the distraction, accepted to see the late patient, rejected the cellular call,
and responded to the switchboard calls. This last one proved the most exciting and
challenging and would be the subject of this case report. A primary care doctor asked
wanted to discuss the case of a young diabetic patient. Taking into consideration
the time pressure, only a short discussion was possible. However, after seeing the
patient who was in the room, he felt that it was better to see the patient himself
on the same day, so he called back the primary care doctor and asked her to send the
patient along to be seen as a walk-in patient.
Case History
A 23-year-old native Emirati single female was diagnosed with diabetes 2 years previously
(2016). She initially presented with classic osmotic symptoms; her initial hemoglobin
A1c (HbA1c) was 10.9%. Despite her young age, typical symptoms, high HbA1c on presentation,
and not particularly morbidly obese, she was assumed to have type 2 diabetes mellitus
(T2DM). She was started on metformin 1g twice daily. Her blood glucose monitoring
values improved reasonably promptly, and her A1c went down to 6.1% within 6 months.
She sustained this level throughout 2017. However, her diabetes started to deteriorate,
reflected in her self-monitoring of blood glucose (SMBG) values and HbA1c (10.1% in
January 2018). Sitagliptin was added in combination with metformin but her HbA1c 3
months later rose to 12%. On May 2, 2018, she changed to gliclazide 60 mg daily, dapagliflozin
10 mg daily, and vildagliptin 50 mg combined with metformin 1,000 mg twice daily as
she remained reluctant to start insulin. For 1 week before the presentation, she had
been feeling vaguely well, tired, and constantly nauseated. She thought it could be
a large number of tablets, and she decided to seek help from her physician, who referred
her to the endocrine clinic.
She confirmed the history as documented above. No other medical problems were found.
There was a family history of T2DM in her mother but no family history of type 1 diabetes
mellitus (T1DM), thyroid disease, or other autoimmune diseases. She looked unwell
and miserable. She was clinically euthyroid with no goiter, no vitiligo, or alopecia.
The unintentional weight loss of more than 10 kg and her HbA1c were noted ([Fig. 1]). The comparison of her random blood glucose and details of her urinalysis on presentation
and now were fascinating ([Table 1]).
Table 1
Comparison of serum blood glucose and urinalysis at the initial presentation in 2016
and the current presentation in 2018
|
Investigations
|
Initial (2016)
|
Current (2018)
|
|
Serum glucose (mmol/L)
|
12.3
|
6.4
|
|
Urine specific gravity
|
1.020
|
1.020
|
|
Urine glucose + (mmol/L)
|
Negative; 0
|
4+ 56
|
|
Urine ketone + (mmol/L)
|
Negative; 0
|
2+ 5.0
|
The dichotomy between plasma glucose and urinary chemistry reflects the stage of diabetes
and the impact of SGLT2 inhibitors.
Fig. 1 The time course of the hemoglobin A1c (HbA1c) (A) and body weight (B) from diagnosis till presentation. The dates on the X-axis apply to both A and B. Minimal variations in the weight measurements at different care facilities are expected.
The blue arrow marks the nadir of the weight corresponding to the initiation of insulin
therapy. Note the U-shape of the HbA1c curve and the dissociation between the two
curves.
At this stage, it became apparent that the diagnosis must be T1DM rather than T2DM.
This was explained to the patient. Investigations to support the diagnosis of T1DM
were requested. However, a complete physiological insulin regimen was initiated immediately
using insulin degludec (Tresiba, Novo Nordisk, Denmark) 10 U daily subcutaneously
and insulin Humalog Lispro (Eli Lilly, United States) 8 U thrice daily. All other
medications were discontinued. The exemption from Ramadan fasting for T1DM for newly
diagnosed people was explained to the patient. She was seen, and arrangements were
made to see her in 1 week jointly with a diabetes educator to review the response
and review the results. Investigations confirmed her low bet cell reserve and autoimmune
background consistent with T1DM ([Table 2]). The patient made a speedy recovery regarding blood glucose and weight ([Fig. 1]).
Table 2
The investigations requested to confirm the diagnosis of T1DM
|
Investigation
|
Patient's results
|
Comments
|
|
Urinary ketones
|
5 mmol/L (2 + )
|
? Starvation or insulin deficiency
|
|
Serum C-peptide
|
0.32 nmol/L
|
Low; normal range: 0.37–1.47
|
|
Serum insulin level
|
6.5 milli IU/L
|
Normal, but does not exclude T1DM
|
|
Anti-GAD antibodies
|
>250 IU/mL
|
Extremely high, normal <4.9
|
Abbreviations: Anti-GAD, anti-glutamic acid decarboxylase autoantibodies; T1DM, type
1 diabetes mellitus.
Serology for celiac disease and autoimmune thyroid disease were negative.
Final Diagnosis and Assessment
-
T1DM presenting with hyperglycemia and no ketosis—obesity proposed the diagnosis of
T2DM.
-
The initial improvement in glycemia is compatible with a “honeymoon” phenomenon rather
than the effect of metformin.
-
The deterioration in blood glucose, weight, and development of ketosis is part of
the natural history of diabetes supported by the lack of response to oral antidiabetic
medications.
-
The average blood glucose on presentation despite high HbA1c and heavy ketonuria is
most likely induced by fasting of Ramadan and the injudicious use of Sodium-Glucose
co-Transporter-2 (SGLT2) inhibitor inferred from the heavy glycosuria. It suggests
that a “euglycemic diabetic ketoacidosis” was imminent if action was taken.
Discussion
When I received this referral, I merely wanted to advise the doctor to start the patient
on basal insulin and send her to the diabetes clinic as soon as possible. Honestly,
I initially only offered to see her on a different day. However, after I cleared all
the mayhem in the clinic, I thought I could have made a significant error of judgment
by leaving the patient in the hands of a doctor who was calling for help and feeling
out of her depth. I do not doubt that such a patient must be assessed by an advanced
care provider, namely specialist diabetes care team. Perhaps the patient's refusal
of insulin therapy could have distracted the physicians from addressing the fundamental
question: what type of diabetes does she have? The answer to this question could have
empowered their arguments in favor of starting insulin as an essential measure rather
than merely one option out of many.[4]
There are a couple of learning opportunities in this case report. Her management deviated
from the principles of good practice for a young patient with newly diagnosed diabetes.[4]
[5] Although admittedly, an attempt was made soon after diagnosis to assess her for
ketonuria, no serious effort was made to explore her autoimmune status and beta-cell
reserve.[1] Her body mass index of 32 kg/m2 on presentation could have made the physicians from further investigations. However,
obesity is prevalent in children, adolescents, and adults in the Gulf, and she was
not exceptionally obese.[6] The family history of T2DM in the mother could have been another red herring. However,
when the patient's glycemic control started to deteriorate, the care providers persevered
with the same treatment plan, and clinical features of T1DM were not sought or ignored.
Indeed, there she was treated as T2DM with no evidence thereof. The risk taken by
ineffective therapies such as sulphonylureas, changing her Dipeptidyl-peptidase 4
(DPP4) inhibitor, or even potentially dangerous treatments such as SGLT2 inhibitors
in a young patient with T1DM has not been excluded.[7]
T1DM is an autoimmune condition resulting from destroying beta cells in the pancreas.
Some patients regain beta cell activity transiently as part of the natural history.
This phenomenon is referred to as the 'honeymoon period' or remission of T1DM.[8]
[9]
[10] This is a very likely scenario for our patient. However, this usually occurs after
a short period of insulin therapy. During this period, patients manifest improved
glycemic control with reduced or no use of insulin or antidiabetic medications.[8]
[9]
[10] Metformin does not cause hypoglycemia, so no need was felt to reduce its dose. However,
the improvement was falsely attributed to metformin, although the magnitude reduction
in HbA1c from greater than 10% down to 6.3% is not typical for metformin action. The
honeymoon period has been more extensively studied in the pediatric population than
the adult population, leading to limited information regarding the honeymoon phase
available to providers of patients with T1DM diagnosed in adulthood. Known factors
predicting remission include high age at onset, male sex, mild initial metabolic derangement,
and absence of frank ketoacidosis.[8]
[9]
[10] Some of these do exist in our patients.
Latent autoimmune diabetes in adults (LADA) can be diagnosed by the cooccurrence of
three traits, adult-onset noninsulin-requiring diabetes, an islet autoantibody such
as glutamic acid decarboxylase autoantibodies or cytoplasmic islet cell autoantibodies,
and no need for insulin treatment for several months postdiagnosis.[1]
[2] In the present patient, over 18 months passed before she presented again with severe
hyperglycemia. This suggests that possible diagnosis of LADA. Similar to T1DM, patients
with LADA often carry other autoantibodies associated with celiac disease and adrenal
and thyroid disorders, but these were not detected in our patients. It is widely agreed
that LADA may represent one end of a continuum of autoimmune diabetes, with classic
T1DM occupying the other end of the spectrum.[2] Interestingly, the mode of the present patient's unintentional weight loss is not
typical of T1DM, that is, it was neither rapid nor associated with poor glycemic control
([Fig. 1]). However, the magnitude of the weight loss should have alerted the physicians to
consider the possibility of T1DM.
Soon after the introduction of this class, the risk of diabetic ketoacidosis (DKA)
after initiation of SGLT2 inhibitors has been suggested by several case reports, including
cases of T1DM.[11]
[12]
[13]
[14] Also, analysis of an extensive database of commercially insured patients[15] suggested that SGLT2 inhibitors were associated with twice the risk of DKA as were
DPP4 inhibitors. However, cases of DKA leading to hospitalization were infrequent.
It has been recommended that the increased risk of DKA with SGLT2 inhibitors is among
the factors to be considered at the time of prescribing and throughout therapy if
patients present with symptoms suggestive of DKA. This was not observed in the present
case. Risks and benefits were not weighted carefully when SGLT2 inhibitors were prescribed.
Two weeks later, the patient was presumably at the prink of euglycemic DKA revealed
in her urinary and plasma chemistry, most likely resulting from the natural progression
of her diabetes, Ramadan fasting, and SGLT2 inhibitors. Fortunately, the patient was
seen and started on insulin in the very nick of time.
Conclusions
The present case highlights the need to consider T1DM and its variants in young patients
with atypical attributes and an unusual course of events at the time of diagnosis
or soon after that. The label of the type of diabetes should be reviewed if the course
changes or when new data become available—injudicious usage of medications that may
be ineffective or potentially carry a risk of harm. Finally, from an organizational
point of view, the management of complex or unusual cases of diabetes remains the
remit of advanced care providers in specialized centers, which should allow a seamless
patient flow at the time of their need.
