**4. Screening and diagnosis**

Dysglycaemia is often detectable for several years prior to a formal diagnosis of CFRD, and early detection is important to prevent pulmonary function decline and weight loss developing [6, 8]. Timely intervention with appropriate treatment may have a profound impact on a patient's health and well-being. In the next section, we will discuss the various methods of screening for CFRD.

### **4.1. Diagnosis**

in part why CF individuals with a familial history of T2DM are at increased risk of developing

It seems likely that genetic modifiers have a significant role in determining to what extent dysglycaemia complicates the clinical course of an individual with cystic fibrosis. [35].

Glucose handling may also be disturbed in individuals with CF-related liver disease, partic‐ ularly in severe cases where the hepatic uptake and storage of glucose is diminished, poten‐ tially rendering treatment for hypoglycaemia with exogenous glucagon ineffective. Chronic kidney disease, which may develop in CF because of recurrent uroliathiasis/nephrocalcinosis, repeated exposure to nephrotoxic medications or indeed as a complication of CFRD itself, may

Although it is unclear how important this is for overall blood sugar control, severe renal failure may lead to an accumulation of exogenous insulin therapy causing an apparent improvement in diabetic control (if judged by insulin requirements) as well as predisposing to hypoglycae‐

Incretin hormones released from the bowel are a major determinant of pancreatic responses in the post-prandial period, causing significantly more insulin to be secreted compared to the presence of glucose alone – this so-called incretin effect is lost in T2DM and a range of treatments that target components of the diffuse endocrine system have revolutionised the treatment of this condition. Whilst it has been shown that enteroendocrine cells are present in the expected amounts in the bowels of both animal models and direct intestinal biopsies from children with CF [44, 45], the incretin effect itself has not been definitively investigated in a CF setting, with various small-scale studies showing conflicting results. Whether or not the incretin system is attenuated in CF, it is unclear whether augmenting its function pharmaco‐ logically would in any way improve the responses of a pathologically damaged pancreas.

Considering the mechanisms above, it is clear to see that although CFRD has a number of features in common with both T1DM and T2DM it is clearly significantly different to both – a

**CFRD T1DM T2DM**

Microvascular Macrovascular

comparison of CFRD with the more common forms of diabetes is shown in Table 1.

Treatment Insulin Insulin Oral agents +/- insulin

Insulin resistance Variable Minimal Significant Insulin deficiency Relative Absolute Variable Typical BMI Low Low High

Ketosis Very rare Common Rare

Complications Pulmonary, weight loss,

**Table 1.** Comparison of CFRD with other forms of diabetes

microvascular

CFRD themselves [35] despite a lack of insulin resistance.

also effect glucose regulation.

92 Cystic Fibrosis in the Light of New Research

mic episodes.

The World Health Organisation, Diabetes UK, American Diabetes Association and UK CF Trust all consider that the formal diagnosis of CFRD should be based on responses to a standard 75 g oral glucose tolerance test (see Table 2) and recommend that it is carried out annually from late childhood. Given the impact of pulmonary decline in early CFRD, tighter thresholds were discussed, but currently the general diagnostic criteria are still applied to CF.

∙ 2-hr 75g OGTT plasma glucose result ≥11.1 mmol/l


**Table 2.** Diagnostic criteria for CFRD (any of the following)

These criteria were originally established by consensus opinion in 1998 [3]. The previously used sub-categorisation of the diabetic response to recognise whether or not fasting hyper‐ glycaemia was also present (often abbreviated as CFRD FH+ and CFRD FH-, respectively), is now considered to be obsolete since the differentiation adds little to management [46].

The OGTT was originally designed to detect individuals, predominantly with insulin resist‐ ance, who were most likely to benefit from intervention aimed at preventing the microvascular sequelae of dysglycaemia. Although these do occur in CFRD, they are relatively late phenom‐ ena, where pulmonary decline (leading to increased mortality) occurs much earlier.

Furthermore, OGTT responses are highly variable in the CF population, where insulin resistance waxes and wanes and glucose absorption alters depending on a variety of factors peculiar to this cohort [17]. An opportunistic annual screen for diabetes carried out when an individual with CF is well may poorly reflect their overall glucose handling.

Screening for diabetes based on the 2-hour glucose value alone has been shown to be of less use in the CF population [47]. Furthermore, the dose of glucose used to stress the endocrine function of the pancreas during an OGTT may be inadequate in CF, where individuals have high calorific requirements and may routinely ingest carbohydrate loads much higher than the 75 g bolus typically used in the diagnostic test.

Given these facts, as well as the typical insulin response profile seen in patients with CF, basing the diagnosis of, as well as screening for, CFRD by means of an OGTT, as is currently recom‐ mended [11], is clearly imperfect [48], leading to the conclusion that the under-diagnosis of clinically relevant dysglycemia is common.

Patients with CF may experience significant hyperglycaemic excursions shortly after eating, which then rapidly correct to the euglycaemic range as the delayed first-phase insulin response belatedly begins; this not only devalues diagnostic tests such as the OGTT (which uses the 2 hour glucose value to diagnose diabetes) but also means that measures of average glucose control, such as glycosylated haemoglobin levels (the HbA1c test), may be misleadingly normal or even low in this group and therefore of much less clinical relevance in CFRD than in other forms of diabetes [49]. One study from Denmark found only 16% of patients had elevated HbA1c values at the time CFRD was diagnosed [50]; others have found similar results [49]. The utility of HbA1c may be further devalued by the increased red cell turnover com‐ monly seen in CF patients [49, 51]. As progressive decline in pancreatic endocrine function in the CF population is related to increased morbidity years before a formal diagnosis of CFRD is made by OGTT [8, 9, 10] and given that HbA1c is at best an unreliable marker of glucose handling [49], other methods of assessing clinically important dysglycaemia, such as 1-hour OGTT result, serial glucose monitoring or Continuous Glucose Monitoring (CGM), have now been advocated as better markers of clinically relevant dysglycaemia.

Elevation of 1-hour glucose level during an OGTT has been shown to be a better predictor of declining pulmonary function than a standard 2-hour result [9] as well as correlating better with other methods of diabetic screening [53]. Currently, the UK CF Trust advocates the use of both a standard 2-hour OGTT as well as a period of serial glucose monitoring (that entails checking levels before and 2-hours after meals and at bedtime for several days) to establish if clinically relevant hyperglycaemia requiring treatment is present [49].

For effective serial glucose monitoring to take place, patients must be empowered and educated with regards to the technique including timing, hand washing, correct sampling procedures as well as meter and strip management. Misleading results have been documented [54] due to poor technique, therefore patient-recorded measurements should be considered with a degree of caution, particularly one-off abnormal readings. Additionally, there is a risk that patients, fearful of a further increase in their treatment burden associated with being diagnosed with CFRD, will report lower glucose readings, fail to document high levels or simply fabricate results altogether.

However, a CGM system circumvents many of the inherent weaknesses associated with selfreported serial monitoring whilst enabling an accurate glucose profile to be established. Worn for a period of 3-5 days typically on the arm or abdomen, a CGM system consists of a trans‐ mitter with a flexible plastic sensor which sits just below the skin allowing regular measure‐ ment of interstitial glucose. Readings are sent wirelessly to a receiver automatically every minute without any user input, although an occasional finger-prick reading is required to calibrate readings against blood glucose levels.

The use of CGM has been validated in the CF population [55] whilst others have shown that it may be superior to other modes of screening for dysglycaemia [56]. The visual output obtained from a CGM can be a useful tool in highlighting sugar control issues with patients

**Figure 3.** CGM in situ with separate base unit Fig 3. CGM in situ with separate base unit

10 15 of treatment, which are as follows: of treatment, which are as follows; Treatment should

12 6 (Humulin M3). (e.g. biphasic isophane insulin).

12 9 basal, bolus basal/bolus

12 7/8 (Humalog mix, Novomix) (e.g. biphasic insulin aspart, lispro)

7 33 HbA1C HbA1c

9 33 at present their occurrence, very

12 1 Intermediate insulin – Onset of action

hours.

8 17 Add fig 3.

8 4 educated with regards to technique educated with regards to the technique

mended [11], is clearly imperfect [48], leading to the conclusion that the under-diagnosis of

Patients with CF may experience significant hyperglycaemic excursions shortly after eating, which then rapidly correct to the euglycaemic range as the delayed first-phase insulin response belatedly begins; this not only devalues diagnostic tests such as the OGTT (which uses the 2 hour glucose value to diagnose diabetes) but also means that measures of average glucose control, such as glycosylated haemoglobin levels (the HbA1c test), may be misleadingly normal or even low in this group and therefore of much less clinical relevance in CFRD than in other forms of diabetes [49]. One study from Denmark found only 16% of patients had elevated HbA1c values at the time CFRD was diagnosed [50]; others have found similar results [49]. The utility of HbA1c may be further devalued by the increased red cell turnover com‐ monly seen in CF patients [49, 51]. As progressive decline in pancreatic endocrine function in the CF population is related to increased morbidity years before a formal diagnosis of CFRD is made by OGTT [8, 9, 10] and given that HbA1c is at best an unreliable marker of glucose handling [49], other methods of assessing clinically important dysglycaemia, such as 1-hour OGTT result, serial glucose monitoring or Continuous Glucose Monitoring (CGM), have now

Elevation of 1-hour glucose level during an OGTT has been shown to be a better predictor of declining pulmonary function than a standard 2-hour result [9] as well as correlating better with other methods of diabetic screening [53]. Currently, the UK CF Trust advocates the use of both a standard 2-hour OGTT as well as a period of serial glucose monitoring (that entails checking levels before and 2-hours after meals and at bedtime for several days) to establish if

For effective serial glucose monitoring to take place, patients must be empowered and educated with regards to the technique including timing, hand washing, correct sampling procedures as well as meter and strip management. Misleading results have been documented [54] due to poor technique, therefore patient-recorded measurements should be considered with a degree of caution, particularly one-off abnormal readings. Additionally, there is a risk that patients, fearful of a further increase in their treatment burden associated with being diagnosed with CFRD, will report lower glucose readings, fail to document high levels or

However, a CGM system circumvents many of the inherent weaknesses associated with selfreported serial monitoring whilst enabling an accurate glucose profile to be established. Worn for a period of 3-5 days typically on the arm or abdomen, a CGM system consists of a trans‐ mitter with a flexible plastic sensor which sits just below the skin allowing regular measure‐ ment of interstitial glucose. Readings are sent wirelessly to a receiver automatically every minute without any user input, although an occasional finger-prick reading is required to

The use of CGM has been validated in the CF population [55] whilst others have shown that it may be superior to other modes of screening for dysglycaemia [56]. The visual output obtained from a CGM can be a useful tool in highlighting sugar control issues with patients

been advocated as better markers of clinically relevant dysglycaemia.

clinically relevant hyperglycaemia requiring treatment is present [49].

simply fabricate results altogether.

calibrate readings against blood glucose levels.

clinically relevant dysglycemia is common.

94 Cystic Fibrosis in the Light of New Research

and allow for highly individualised treatment regimens to be decided. A positive CGM result – that is a value of more than 7.87.8mmol/l for over 4.5% of the test – is associated with a worse prognosis in terms of CF-specific endpoints (i.e. pulmonary function and weight) as well as being a better predictor of future risk of development of CFRD than an OGTT [57], which has been shown to give unreliable results when used as a screening tool [58]. See fig 3: 8 22 7.8 7.8mmol/l 9 23 (43) [43]

at present their occurrence very infrequently

#### **5. Clinical features, signs and symptoms** infrequently, complicates complicates

The biochemical and clinical detection and diagnosis of CFRD can be difficult, since symptoms of polyuria, polydipsia and weight loss only occur in one third of cases [8]. An unexplained decline in pulmonary function or difficulty maintaining or gaining weight should alert health care professionals to the possibility of underlying significant dysglycaemia, impaired glucose tolerance or even CFRD. Poor weight gain or weight loss can often be identified in the preceding 12 months prior to a diagnosis of CFRD [57]. be considered: 11 6 diagnosing managing 11 34 control levels 11 37 (Lantus, Levemir, Tresiba) (e.g. insulin glargine; detemir or degludec)

#### **6. Consequences of dysglycaemia in CF** 2-3 hours with a duration of 16-35 a duration of 16-35 hours (e.g. insulin isophane)

Multiple studies and reviews of large registry datasets have demonstrated a clear link between dysglycaemia and poor outcomes in pulmonary health in CF [3, 5, 43], but the mechanism by which hyperglycaemia, often intermittent rather than sustained in this population, causes the 12 2/3 (Actrapid,Humilin S) (e.g. soluble insulin) 12 5 (Novorapid, Humalog, Apidra). (e.g. insulin aspart, lispro or glulisine)

Intermediate insulin – Onset of action 2-3 hours with

2

observed decline in lung function is less clear. Plausible theories include direct pulmonary damage or suppression of various components of the immune system in the setting of hyperglycaemia leading to an increased risk of pulmonary infections which might also be more severe in nature [14], as has been demonstrated in other types of diabetes [44]. Insulin therapy has been shown to reduce sputum pathogen levels in CF patients [45]. It has also been hypothesised that the predominant factor in the decline in both respiratory function and nutritional status is a direct consequence of the altered protein metabolism experienced by CF patients who become relatively deficient in insulin therefore losing its important anabolic effects [14]. CFRD has been considered to be a risk factor for the development of distal intestinal obstruction syndrome in cystic fibrosis patients [59]; however, it is unclear how strong the association really is between these two conditions [60] or what causal mechanism might be involved.
