Metformin Discontinuation for 48 Hours Reduces Intestinal Fluorodeoxyglucose Uptake in 18F-fluorodeoxyglucose Positron Emission Tomography/Computed Tomography

Hong Kong J Radiol 2021 Sep;24(3):186-91

 

 

Nuclear Medicine Unit and Clinical PET Centre, Queen Elizabeth Hospital, Hong Kong

 

  Full Article in PDF

Metformin is an oral antihyperglycaemic agent commonly used in non-insulin dependent (type II) diabetes mellitus. It is known to be associated with typically intense, diffuse, and continuous uptake along the bowel in ¹⁸F-fluorodeoxyglucose positron emission tomography and computed tomography (FDG PET/CT).[1] This can pose difficulty in FDG PET/CT interpretation by obscuring bowel lesions or adjacent extra-intestinal structures. In FDG PET/CT, a particular concern of metformin discontinuation is the effect on blood glucose levels, which, if >11 mmol/L,[2] may delay FDG PET/CT scanning.

 

Previous studies on metformin discontinuation in non-Chinese populations had different study designs and showed conflicting results in the effect of metformin discontinuation on examination day blood glucose levels.[3] [4] The purpose of this study was to evaluate the effect of 48-hour metformin discontinuation on reducing bowel FDG uptake in a Chinese population undergoing FDG PET/CT, and the effect on blood glucose levels.

 

This was a prospective single-centre study with all patients recruited from January 2019 to July 2019 who were referred to our PET/CT centre for ongoing assessment of various neoplastic conditions. Inclusion criteria were patients with type II diabetes being treated with metformin as monotherapy or as part of polydrug therapy, and who had undergone a previous FDG PET/CT in our centre within the past year.

 

Exclusion criteria were patients with known gastrointestinal malignancy, previous bowel resection, or other intra-abdominal pathology that might affect bowel uptake, e.g., an enterovesical fistula that might cause the appearance of urinary excretion of FDG into the bowel (Table 1).

 

Table 1. Excluded conditions (n = 8) and frequency of occurrence.

 

Patients were instructed to stop taking metformin for 48 hours prior to undergoing the second FDG PET/CT study in our centre. Patients were interviewed on arrival to PET/CT centre by nursing staff to confirm metformin stoppage. Patients were also reminded to resume metformin after PET/CT study was completed.

 

Patients’ blood glucose levels were checked and FDG was only injected if they measured <11 mmol/L, in accordance with local protocol and international guidelines.[3] Other parameters, including body weight, age, sex, metformin daily dosage, injected ¹⁸F-FDG activity, and uptake time were recorded for each patient.

 

The two PET/CT studies were performed in each patient using the same integrated PET/CT scanner (Discovery 710, General Electric, Milwaukee [WI], United States) at Queen Elizabeth Hospital, Hong Kong. Patients were instructed to fast for at least 6 hours before ¹⁸F-FDG injection. Blood glucose levels were measured and the FDG PET/CT was only performed if blood glucose level was <11 mmol/L. Patients were injected with 370 MBq (for normal-weight patients) or 555 MBq (for patients with body weight >80 kg) according to department protocol. Actual dose administered ranged from 347 to 609 MBq, as measured by dose calibrator. Acquisition commenced 60 minutes after FDG administration. CT scanning for anatomical localisation and attenuation correction was performed with the following parameters: 120-kV tube voltage, 120-mA tube current, 0.5-s gantry rotation time, and 0.984 pitch. The PET images were acquired in three-dimensional mode, from skull to mid-thigh with 2 minutes for each bed position. The PET raw data were processed using ordered subset expectation maximisation, point spread function modelling, and time-of-flight (four iterations with 18 subsets and 5.5-mm cut-off frequency). The data were reconstructed with 3.75-mm section thickness in a 256- × 256-mm matrix and processed through a standard filter.

 

Attenuation-corrected PET images, maximal intensity projection images, CT images, and PET/CT fused images were generated and displayed using a dedicated workstation, AW VolumeShare 7 (AW 4.7 Ext. 8 Software, General Electric). The FDG uptake was graded both visually and semiquantitatively over different bowel segments by an observer with 5 years of experience in PET/CT.

 

A four-point score scale described by Gontier et al[1] was used for visual assessment of the small and large bowel FDG uptake: Grade 1 (lower than hepatic activity); Grade 2 (similar to hepatic activity); Grade 3 (moderately higher than hepatic activity); and Grade 4 (diffuse and intense uptake).

 

Semiquantitative analysis of the FDG uptake over different bowel segments was performed using maximum standardised uptake value (SUVmax). The SUVmax was measured using a region of interest of 1 cm in diameter over the predefined anatomical locations, including: horizontal portion of duodenum using the pancreatic head as landmark; jejunum measured at mid-height of descending colon; distal ileum adjacent to the ileocaecal valve; caecum; hepatic flexure; splenic flexure; and junction between descending colon and sigmoid colon.

 

Statistical analysis was performed using SPSS (Windows version 22.0; IBM Corp, Armonk [NY], United States). A paired-sample t test was used for comparison of visual analysis scores and SUVmax measurements between the two successive PET/CT studies. A p value <0.05 was considered significant.

 

In total, 44 patients were recruited. Two patients who stopped metformin only for 24 hours and one patient who did not stop taking metformin in the second PET/CT study were excluded. A total of 41 patients (23 men, 18 women) with mean age of 67.2 ± 10.7 years were included in the final analysis. Patients’ parameters including weight, body mass index, and metformin daily dose between the two scans were unchanged. The injected doses (402.1 MBq ± 54.1 vs. 409.7 MBq ± 51.7; p = 0.299) and time between injection and imaging between the two scans showed no statistically significant difference (Table 2). Effect of Metformin

 

Table 2. Summary of patient and imaging parameters.

 

Small and large bowel FDG uptakes were significantly reduced in the second FDG PET/CT study with metformin stoppage for 48 hours, measured by visual scoring and SUVmax. Significant differences in visual scoring of the bowel uptake were found in small bowel (3.10 ± 1.00 vs. 1.63 ± 0.73) and large bowel (3.61 ± 0.54 vs. 2.54 ± 0.90). For semiquantitative analysis, there was a significantly lower SUVmax over all the predefined bowel segments except at the duodenum (Table 3, Figure 1).

 

Table 3. Results of the semiquantitative analysis (SUVmax) of bowel uptakes in the two consecutive PET/CT studies before and after stopping metformin.

 

Figure 1. Histogram plot showing mean visual analysis scores between the two successive FDG PET/CT examinations. Error bars indicate standard deviations. Asterisks indicate statistical significance.

 

Metformin daily dosage among the patients recruited ranged from 500 to 2000 mg/d and were not changed between the two PET/CT studies. A mean increase in examination day blood glucose level of 1.07 mmol/L was found when metformin was withheld for 48 hours. All patients in this study had blood glucose levels <11 mmol/L after discontinuation of metformin and did not require injection of short-acting insulin.

 

In a patient with history of renal cell carcinoma, there was focal uptake in the caecum in the second FDG PET/CT study performed after metformin discontinuation (Figure 2). This focal uptake was not discernible in the first FDG PET/CT with the presence of metformin-related intense bowel activity. The large bowel uptakes were scored Grade 4 and Grade 1, respectively, in the two consecutive PET/CT studies according to the visual analysis score.

 

Figure 2. A case of renal cell carcinoma of the right kidney with lung metastases in a 64-year-old man was referred for follow-up FDG PET/CT. MIP image of the FDG PET-CT study on metformin (a) showing intense small and large bowel activity of grade 4 by the visual score scale. MIP image of the second FDG PET-CT study after metformin discontinuation showed focal uptake in the caecum of SUVmax 6.7, now discernible with reduced large bowel uptake of grade 1. The patient was advised to undergo colonoscopy for evaluation.

 

The exact mechanism of how metformin affects intestinal glucose uptake and, hence, bowel FDG uptake, is not entirely known. Ethnic differences in response to metformin therapy have been reported.[5] [6] [7] Thus a local population study to evaluate the effect of metformin on bowel FDG uptake is warranted before implementing changes in our imaging protocol. To the best of our knowledge, this is the first study in a Chinese population that examined the effect of metformin discontinuation on intestinal FDG uptake. Our study showed that metformin discontinuation for 48 hours is effective in reducing bowel FDG uptake in our local population with results comparable to those of prior studies.[8] [9]

 

In FDG PET/CT study a practical consideration in suspending metformin is the potential effect on blood glucose levels on the day of the examination. According to the European Association of Nuclear Medicine procedure guidelines for tumour imaging, a FDG PET/CT study should be delayed or rescheduled if the blood glucose level is >11 mmol/L. This would create problems in PET centre scheduling and, more importantly, a delay in diagnosis. Contrary to previous study[6] which asked patients to stop all antihyperglycaemic drugs, we specifically asked patients to suspend only metformin, whilst continuing other glucose-lowering medication(s). Slightly higher examination day blood glucose levels were observed with metformin discontinuation, but all patients had blood glucose levels <11 mmol/L, thus allowing timely performance of FDG PET/CT studies.

 

One merit of performing FDG PET/CT after metformin discontinuation is that the reduced bowel uptake facilitates potential bowel lesion detection, as illustrated by a case in our study where an incidental focal colonic uptake in the caecum was only detected in the second FDG PET/CT performed after metformin discontinuation. The identification of focal colonic uptake is of important clinical significance in FDG PET/CT, as the reported pooled risk of malignant or pre-malignant lesions appearing as focal colonic uptake on FDG PET/CT was 68% in a meta-analysis[9] and the authors stated that further investigation is warranted whenever focal colonic uptake is detected. Similar findings regarding focal colonic uptake and underlying colonic polypoid lesion detection in FDG PET/CT have also been reported in our centre, with a prevalence of focal colonic uptake of 4.8%, comparable to previous studies.[10] [11] It has been suggested that increased bowel activity (SUVmax >5.9) encountered without discontinuing metformin would obscure focal colonic uptake detection and hinder appropriate management, such as colonoscopy or virtual colonoscopy according to an evidence-based review.[12] Further study is warranted to evaluate the true incidence of focal colonic uptake after metformin discontinuation.

 

There are limitations to our study. First, this was a single-centre study with a relatively limited number of patients recruited. Second, it was a single-observer non-blinded study, making it prone to observer bias. However, we adopted the visual scale taking the liver uptake as an internal reference, which is relatively consistent even in the presence of diffuse liver disease.[13] [14] [15] We also predefined anatomical localisations of bowel segments to reduce sampling bias and achieve higher reproducibility.

 

In summary, discontinuation of metformin for 48 hours prior to FDG injection and scanning significantly reduced bowel FDG uptake, which may facilitate bowel lesion detection. It is an effective and feasible preparation for FDG PET/CT studies and should be considered in diabetic patients treated with metformin.