Honeymoon phase in type 1 diabetes mellitus: A window of opportunity for diabetes reversal?

Abstract

The knowledge of the pathogenesis of type 1 diabetes mellitus (T1DM) continues to rapidly evolve. The natural course of the disease can be described in four clinical stages based on the autoimmune markers and glycemic status. Not all individuals of T1DM progress in that specific sequence. We hereby present a case of T1DM with a classical third phase (honeymoon phase) and discuss the intricacies of this interesting phase along with a possible future promise of “cure” with the use of immunotherapies. We now know that the course of T1DM may not be in only one direction towards further progression; rather the disease may have a waxing and waning course with even reversal of type 1 diabetes concept being discussed. The third phase popularly called the “honeymoon phase”, is of special interest as this phase is complex in its pathogenesis. The honeymoon phase of T1DM seems to provide the best window of opportunity for using targeted therapies using various immunomodulatory agents leading to the possibility of achieving the elusive “diabetes reversal” in T1DM. Identifying this phase is therefore the key, with a lot of varying criteria having been proposed.

Key Words: Diabetes reversal; Diabetes remission; Type 1 diabetes; Honeymoon phase; Immunomodulation

Core Tip: The insulin requirement reduces drastically in the Honeymoon phase or the partial remission phase in type 1 diabetes mellitus (T1DM). Novel pharmacological interventions and immunomodulating therapies in T1DM are being tried specifically in this phase as it may serve as the potential window of opportunity for possible “cure”. It may therefore be clinically relevant to timely identify this phase in these patients.

INTRODUCTION

Type 1 diabetes mellitus

Type 1 diabetes mellitus (T1DM), previously known as Insulin Dependent Diabetes, results from the autoimmune destruction of beta cells of pancreatic islets. With genetic susceptibility conferring the risk of diabetes mellitus since birth, the occurrence of autoantibodies against beta cell components like insulin, Islet-specific Glucose-6-phosphatase catalytic subunit-related protein, Zinc transporter 8, Glutamic Acid Decarboxylase (GAD), Islet Antigen-2 (IA-2) contribute to the loss of immunologic tolerance to beta cells and cause their progressive depletion. Interesting to note is that the mere presence of antibodies doesn’t lead to diabetes, emphasizing the role of other factors like environment as well as stressed beta cell dysfunction in their functional decline.

Stages and phases of T1DM

The natural course of T1DM can be described in four clinical stages based on the autoimmune markers and glycemic status. Stage 1 is characterized by autoantibody positivity which predisposes the individual to autoimmune destruction of pancreatic beta cells thereby leading to insulin deficiency and metabolic dysfunction which constitutes Stage 2. Stage 3 denotes overt hyperglycemia which is classically defined as the T1DM onset. Stage 4 indicates the post-diagnosis period. These stages although assist in understanding the disease, it is to be borne in mind that not all individuals of T1DM progress in that specific sequence. Similarly, T1DM is also described in four phases namely pre-clinical, overt, partial remission, and chronic phase, of which the third phase popularly called the honeymoon phase has instilled extreme interest. This phase is complex in its pathogenesis as well as intervening in this stage seems to be an extremely feasible option for the concept of T1DM reversal. Here we discuss the various criteria for diagnosing the honeymoon phase and present case details of a patient in this phase and the various factors affecting it.

CASE PRESENTATION

An 11-year-old boy, second born to non-consanguineously married parents, had osmotic symptoms, generalized weakness for 4 wk, and hyperglycemia (> 500 mg/dL). At presentation, he had persisting polyuria with no features of hypothyroidism/adrenal insufficiency/celiac disease. On examination, he was not dehydrated, with height above 97^th centile for age and weight between 50-75^th centile. The child had a pubertal status of P2 G2 as per Tanner staging. Blood investigations revealed a raised anion gap with mildly positive urine ketones, and no acidosis with normal serum electrolytes. Kidney function, liver function tests, and thyroid profile were within normal limits and HbA1c was > 14.5%. He was initially managed with intravenous fluids, insulin infusion with potassium supplementation as needed and later shifted to subcutaneous basal-bolus insulin regimen. On evaluation, serum C-peptide was 0.34 ng/mL, serum insulin was 1.106 mIU/L, anti-TPO (53.5 IU/mL) and anti-GAD 65 antibodies (208 IU/mL) were positive with other antibodies including IA2 and anti-insulin antibodies being negative. He was discharged on the basal-bolus regimen with a total daily insulin requirement of 48 units/d (1.1 units/kg/d). The patient was followed up closely on an outpatient basis and his insulin requirements gradually decreased when at 8 wk the insulin requirement was only 2-4 units/d (0.1 unit/kg/d) with HbA1C of 7.6%. This minimal insulin requirement persisted for nearly a year with HbA1c values maintaining near 6.5%, suggesting a phase of partial remission.

PARTIAL REMISSION - DEFINITIONS AND CRITERIA

The phase of partial remission (also known as Honeymoon phase) has revived interest owing to the possibility of pharmacological intervention and immunomodulation. This phase is clinically characterized by decreasing (partial) or rarely absent (complete remission) insulin requirements and near-normal glycemic control in patients of type 1 DM. The definitions of this phase have evolved much since Jackson first described it in the 1940s[1]. As in T1DM, the insulin requirement was around 0.8-1.2 units/kg body weight/24 h, authors initially labeled this phase as insulin requirement of < 0.5 units/kg body weight/24 h with few even suggesting a lower dose of 0.3 units/kg body weight/24 h[2,3]. Insulin requirement has been used as a surrogate measure of stimulated C-peptide. The less commonly used definition of decrease in insulin dose of at least 33% from dose at diagnosis has also been employed in a few studies[4]. Achievement of near-normal HbA1C which reflects glycemic control has also been added as a criterion[5]. So, combining both, this phase has classically been defined where insulin requirement is < 0.5 units/kg body weight/ day and HbA1c < 7%[6]. A newer formula proposed by Mortensen et al[7], “Insulin Dose-Adjusted HbA1c” (IDAA1c) is defined as HbA1C+ 4 × (insulin dose in units/kg/24 h) whose value when less than or equal to 9 would indicate partial remission. This definition has been validated well in larger cohorts and correlates with functional beta cell mass and stimulated C-peptide, hence avoiding the need for measurement of stimulated C-peptide[8]. In our patient the IDAA1c value was 8. Another recent retrospective study has claimed a definition of Glycemic Target Adjusted HbA1c which is calculated as HbA1c (%) - [3 × % of normoglycemic values (70-180)] where a value of < 4.5% correlated well with IDAA1c and had a sensitivity and specificity of 72% and 92% in predicting partial remission[9]. Complete remission is defined as insulin independence with HbA1c < 6% on oral or no hypoglycemic agents. The different definitions for partial remission are summarized in Table 1.

Table 1 Different criteria for defining partial remission phase.

Definitions of partial remission phase
Muhammad et al[3]	Insulin requirement < 0.5 units/kg body weight/24 h
Bonfanti et al[2]	Insulin requirement < 0.3 units/kg body weight/24 h
Jefferson et al[4]	Fall in insulin dose of at least 33% from dose at diagnosis
Chase et al[5]	Achieving near normal HbA1C
Araszkiewicz et al[20]	Insulin requirement < 0.3 units/kg body weight/24 h + C-peptide > 0.5 ng/mL + normal 24-h glycemic profile
Mortensen et al[7]	Insulin Dose- Adjusted HbA1C < 9
ISPAD[6]	Insulin requirement < 0.5 units/kg body weight/24 h and HbA1C < 7%
Nielens et al[9]	Glycemic Target Adjusted HbA1C (GTAA1c) < 4.5%

ISPAD: International Society for Pediatric and Adolescent Diabetes.

FACTORS AFFECTING PARTIAL REMISSION

Factors like gender, age at diagnosis of diabetes, body mass index (BMI), duration of diabetes, degree of metabolic compensation, C-peptide levels, HbA1c, antibody positivity, and glycemic control achieved with treatment determine the attainment of partial remission (Figure 1). Boys attain partial remission more frequently than girls whereas in few studies, gender had no association with partial remission occurrence[8,10,11].

Figure 1 Factors predicting partial remission in type 1 diabetes mellitus. DKA: Diabetic ketoacidosis.

In our case, the child was 11 years at diagnosis, male gender, BMI 14.89 kg/m², 3 mo duration of diabetes, history of mild diabetic ketoacidosis (DKA), C-peptide level 0.34 ng/mL, HbA1c > 14.5% at diagnosis, GAD antibody positive, and very good glycemic control was achieved with multiple subcutaneous insulin injections in 3 mo.

Studies have shown that partial remission was more frequent when the age at diagnosis was above 15 years i.e. post-pubertal, while remission is never attained in patients diagnosed under the age of 2 years[2,3,11]. Few studies have also shown the highest rates of remission occur in children aged 5-9 years[12]. The difference in frequencies among different age groups could be explained by the difference in treatment approach varying with age as a study by Chobot et al[10] showed that there was no association of occurrence of remission with age at the onset. In the guideline followed in their study, treatment was uniform in all age groups.

The study by Chobot et al[10] showed that patients who attained remission had a higher BMI compared to those who failed to attain remission. The possible explanation hypothesized is that higher BMI accelerates insulin resistance and worsening glycemic control which in turn aggravates beta cell apoptosis, thus patients present early and gain therapy. This study also revealed that individuals with another concomitant autoimmune disease had more frequent remission which is explained by the possibility of earlier diagnosis.

Ethnic differences seem to have a role in remission occurrence with studies showing lower remission rates among African-Americans and Hispanic children compared to European children. Environmental factors, genetics, epigenetic modifications, and possible resource inequalities could explain the ethnic variations in remission[3].

The absence of HLA DR3 or HLA DR4 and HLA DQB1*0201 have a lower likelihood of remission as shown by a few studies as these genetic polymorphisms were associated with higher C-peptide levels and beta cell concentration. Lower HbA1C levels at diagnosis showed increased chances of remission in a few studies[3,10]. Similarly, higher C-peptide levels at diagnosis favored higher rates of remission in a few studies[12] while others showed no such association[10,12].

The degree of decompensation in terms of occurrence of DKA at the onset of disease has a role in determining remission. It is believed that when patients present with DKA at onset, it corresponds to lesser remaining functional beta cells, thereby in such cases, the remission chance is also negligible[3,11,12]. Likewise, remitters have higher blood pH at diagnosis than non-remitters[10]. Presence of lesser classical symptoms at diagnosis also predicts partial remission also explainable by the above said concept. Similarly, a recent history of infection before diagnosis also predicts lesser chances of remission as infection increases the insulin requirement.

Antibody negativity at diagnosis predicts a higher chance of remission[8]. Initiating intensive glycemic control from the time of diagnosis either by insulin infusion or insulin pump favors remission as explained by “Beta cell rest theory”. According to this theory, providing exogenous insulin decreases endogenous production and provides beta cell rest which allows decreased antigen exposure and halts further autoimmune destruction. Decreased glucotoxicity on beta cells due to intensive control is an added benefit.

Interestingly, ketosis-prone diabetes (KPD) in a black, middle aged, overweight, or mildly obese, black adult male would point to an entity like Flatbush diabetes[13]. The clinical course of KPD is more like that of type 2 diabetes (controlled on diet with/without oral medications after initial intensive insulin therapy) rather than of patients with type 1 diabetes and such patients may be completely off insulin in the interim. Our patient was a 11-year non-obese male child of South Asian origin with GAD antibody positivity.

DURATION AND COURSE

The percentage of T1DM patients attaining remission ranges from 18% to 72% whereas complete remission is extremely rare occurring in less than 3%[11]. The partial remission phase occurs around 3 mo from initiation of therapy in most studies although it may range from 3 to 12 mo[3,10,11]. Studies quote that 35%-70% of remission occurs within 3 mo of diagnosis and the rate of remission progressively declines with increasing duration of disease[8]. The duration of remission ranges from 1 month to 13 years with a mean of 7 mo[10]. Similar to factors described in the diabetes risk[14], the duration of partial remission is also affected by multiple factors. Remission is longer in those with one antibody positivity, older age of onset of diabetes, higher blood pH at diagnosis, and in boys[15,10,11]. In addition to the factors described above, others like inflammatory markers, diet, and nicotinamide are some factors which may also have a role in predicting the occurrence of remission[16]. Flowchart to detect partial remission phase in type 1 diabetes mellitus similar to a simplified model suggested for retinopathy detection in diabetic patients by Johora et al[17] (Figure 2).

Figure 2 Flowchart to detect and diagnose partial remission in type 1 diabetes mellitus. T1DM: Type 1 diabetes mellitus.

PATHOPHYSIOLOGICAL MECHANISMS:

The potential explanations for this third phase of the disease have been multiple. Immunological tolerance to beta cell antigens, enhanced insulin sensitivity, beta cell regeneration or recovery, and satisfactory metabolic control by insulin therapy reversing the effects of glucose toxicity have been the most plausible theories. The insulin resistance in peripheral tissues exists immediately post ketoacidosis resolution and then improves with insulin administration and this enhances insulin sensitivity which correlates with the partial remission phase. With the accepted fact of the autoimmune process with near-total destruction of beta cells leading to disease manifestation, it is also to be considered that the concurrent beta-cell turnover with net regeneration and recovery of residual cells play a significant part in remission. It has also been emphasized that hyperglycemia directly attenuates pancreatic cells’ secretory response as well as insulin-mediated glucose transport, leading to a vicious cycle of hyperglycemia and glucotoxicity which when appropriately halted by satisfactory metabolic control may transiently restore the beta cell function[18].

CLINICAL SIGNIFICANCE

Patients with a history of partial remission in the past have better metabolic control later in the disease course in terms of lower HbA1C and better lipid profiles as well as reduced risk of chronic microvascular complications[10,11]. It is also seen that the hypoglycemic events are lower compared to those with no remission, because the residual b-cell assists in the maintenance of appropriate alpha cells’ glucagon secretion to avoid hypoglycemia[19].

IMMUNOTHERAPY

Although various therapeutic options have been tried in different stages of diabetes with the agenda of primary or secondary prevention, the overall curiosity of the phase of partial remission rests on the predicted prospect of attempting a reversal of diabetes if intervened well in time. With autoimmunity as the main etiology in the pathogenesis of T1DM, drugs like prednisone, cyclosporine, and azathioprine which are non-specific in immunosuppression have been used. Although proving some efficacy, the concerns of long-term drug use, toxicity, and lifelong immunosuppression make their use impractical. For innate immunity, agents like etanercept and canakinumab targeting Tumor Necrosis Factor-alpha and interleukin-1-β respectively have been tried, of which the latter has failed to show any benefit. Furthermore, anti-IL-6 receptor monoclonal antibodies and anti-IL-12/23 monoclonal antibodies are also being evaluated. The role of adaptive immunity has been exploited for the use of anti-CD-20 monoclonal antibody rituximab and anti-CD3 antibody teplizumab in new-onset T1DM. T-cell modulating agents have also been studied in the quest for type 1 diabetes reversal.

CONCLUSION

With multiple existing criteria posing a challenge in diagnosis, rightly picking up the Honeymoon phase or the partial remission phase may serve as the potential window of opportunity for using novel pharmacological interventions and immunomodulating therapies to attain the elusive “Diabetes Reversal” in T1DM. Targeted immunotherapies hold the promise of a possible “cure” in this disease with complex autoimmune-mediated pathogenesis and can thus influence long-term outcomes. Identifying this phase of partial remission is thus a key for using immunomodulatory therapies in T1DM.