Is There Any Need for Adjusting ¹³¹I Activity for the Treatment of High Turnover Graves’ Disease Compared to Normal Turnover Patients? Results from a Retrospective Cohort Study Validated by Propensity Score Analysis

Abstract

Purpose

To compare ¹³¹I-therapy outcomes in high turnover and normal turnover Graves’ disease patients and predict optimal first ¹³¹I activity for high turnover patients.

Methods

Retrospective cohort design (1:2) validated by propensity score analysis. Cohort 1, high turnover (2-h RAIU/24-h RAIU ≥ 1), n = 104, and cohort 2, normal turnover (ratio < 1), n = 208, patients were compared for post ¹³¹I outcome. The cure was defined as a combined euthyroid and stable hypothyroid state following ¹³¹I treatment. Logistic regression analysis was used for identifying prognostic factors. The propensity score was applied; 77 matched pairs (1:1 ratio) of high and normal turnover patients were selected as a validation set.

Results

First ¹³¹I cure rates of 28% in high turnover and 66% in normal turnover groups (p = 0.001) were noted. The therapy cycles (median, 2 vs. 1) and cumulative ¹³¹I activity (median, 15 vs. 7 mCi) were required to cure hyperthyroidism in cohort 1 and cohort 2, respectively. Age (> 44 years), higher grade of goitre, and 2-h RAIU (> 37%) were associated with ¹³¹I therapy failure. The high turnover patients needed a factor of 1.5–2 times more ¹³¹I activity to achieve a similar cure rate compared to the normal turnover patients. The first-dose cure rate was 31% vs. 60% by propensity score analysis (n = 154), no way different (28% vs.66%) from the whole group of 312 patients.

Conclusion

High turnover Graves’ disease patients, if administered standard ¹³¹I activity, the outcomes shall be poor. To improve the success rate, ¹³¹I activity should be increased by 1.5 to 2 times in the high turnover patients.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13139-020-00674-3.

Introduction

Graves’ disease (GD) can be managed with antithyroid drugs (ATD), ¹³¹I radioiodine therapy (RIT), and surgery. All three treatment options are useful, though some consider choosing one option over the other [1–3]. Regarding RIT, the ¹³¹I activity to be administered can be planned by different methods, including administering fixed activity of ¹³¹Ior adjusting ¹³¹I activity based on the size of the gland and gland’s ability to trap radioactive iodine or a more complex method involving detailed pre-therapy dosimetry, requiring calculations of residence time (effective half-life) of ^13II in the thyroid gland [4–11]. Regarding kinetics of ¹³¹I in the thyroid, around15% of GD patients can have a rapid turnover of ¹³¹I in the gland, and the effective half-life (T_eff) for ¹³¹I clearance in this subgroup is significantly less [12–17]. The lower residence time (effective half-life) of ¹³¹I results in lesser radiation absorbed dose delivered to the thyroid gland (target organ), which probably leads to treatment failure in high turnover GD patients if routine activities of ¹³¹I are administered. For this subgroup, various regimens are tried, including the use of higher ¹³¹I activities, some supporting concurrent use of lithium, and some showing no benefit of adding lithium [18–21]. The optimal ¹³¹I activity used in the high turnover subgroup of GD patients is not well defined [22, 23]. With this background, this retrospective cohort study was planned. However, the retrospective cohort studies are fraught with a high degree of bias. To overcome the bias issue, the propensity score, a statistical technique, is applied to create unbiased data set by choosing matched pair for analysis.

Objectives

To compare the outcome of ¹³¹I RIT in high turnover and normal turnover GD patients and predict the optimal first ¹³¹I RIT activity to be used in high turnover GD patients. Results from the retrospective cohort study shall be verified by propensity score analysis.

Method

Study Design

Retrospective cohort design validated by propensity score analysis.

Setting

The institute ethics committee (Human studies), All India Institute of Medical Sciences, New Delhi, approved this retrospective cohort study (ref no: IECPG-598/24.10.2019). The patients were derived from a review of medical records of 5714 consecutive hyperthyroid patients who visited the Thyroid Clinic of Department of Nuclear Medicine, AIIMS, New Delhi, from February 1989 to August 2018 for ¹³¹I RIT.

Inclusion Criteria

GD patients with available baseline 2-h RAIU and 24-h RAIU values before RIT (n = 4800) and at least 1 year follow-up after RIT were included.

Exclusion Criteria

Patients with hyperthyroidism diagnosis other than Graves’ disease (toxic MNG and toxic adenomas) and the ones with incomplete follow-up at Thyroid Clinic (because of convenience, patients went back to the referring physicians/endocrinologists) after ¹³¹I RIT were excluded.

Definitions of Patient’s Groups (Cohorts) Based on the Turnover of the Thyroid Gland

Ratios of early RAIU to late RAIU values were taken as markers for the iodine turnover of the gland [13]. Although different centres use different time points (ranging from 2 to 6 h) for early RAIU [13, 24], as per our institutional practice, we used 2 h for early RAIU [25].

• High turnover—defined as 2-h RAIU/24-h RAIU ratio ≥ 1

• Normal turnover—defined as 2-h RAIU/24-h RAIU ratio < 1

Sample Size

Sample size estimation was done considering the primary objective, which is comparing outcome proportions. We used a formula for 1:2 group allocations for sample size [26, 27]. Considering significance criterion 0.05 (95%) (Z_{crt α/2}:1.96), statistical power 95% (Z_{pwr, β}: 1.645), estimated P₁ as 60%, and estimated P₂ as 80%, the estimated sample size required was 102 patients in the test group [13, 20, 28].

The patient selection chart is shown in Fig. 1. A total of 5714 hyperthyroid patients record were available for review. Out of these, 4800 patients had 2-h RAIU and 24-h RAIU values available. Among these, 199 patients had high turnover (2-h RAIU/24-h RAIU ratio ≥ 1). Out of these 199, eight patients on a thorough review of the case chart had a diagnosis of non-GD, and 87 were excluded as the follow-up was less than 1 year. Therefore, finally, 104 patients for high turnover cohort 1 were identified. For selecting the normal turnover cohort, which was age and sex-matched to a high turnover cohort, we first grouped all the case record into male and female groups and then further stratified all patients in the age group of + 1 year, for example, 19 + 1 (18–20) year, 22 + 1(21–23) year, and 25 + 1 (24–26) year. And then, for each high turnover patient in each slot, we selected 2 normal turnover patients randomly from the same strata resulting in 1:2 age and sex-matched groups. Finally, we included 104 patients in the high turnover group and 208 (1:2 ratio) in normal turnover as a control group.

Fig. 1.

Flowchart showing a selection of patient cohorts, follow-up, and outcome in high and normal turnover groups. Cohort 1 of 104 Graves’ disease patients with high turnover and cohort 2 as the control group (1:2 age/sex-matched normal turnover GD) were selected

Clinical Data

The patient’s data on baseline demographics, namely, thyroid function tests, number of ¹³¹I RIT given, and cumulative ¹³¹I activity administered, were extracted from the treatment chart.

For grading goitre severity, the World Health Organization (WHO) grading of goitre was used, which is described as grade 0, no goitre and not visible nor palpable; grade 1, palpable not visible gland; and grade 2, neck swelling visible when the neck is in normal position, corresponding to enlarged thyroid swelling on palpation [29]. For grading eye proptosis, grade 0, normal; grade 1, only stare present; and grade 2, proptosis present, were used.

¹³¹I Radioiodine Therapy and Follow-Up

Patients were per orally administered ¹³¹I activity ranging from 5 to 15 mCi. Radiation safety instructions were thoroughly explained, and written instructions were handed over to the patients to follow at home strictly. Normally antithyroid drugs are not prescribed routinely at our institution unless a severe comorbid condition was associated with GD. However, beta-blockers were prescribed to all if there were no contra-indication for the same for 6–8 weeks following RIT. Patients were followed at 3 monthly intervals with clinical and biochemical assessments at each follow-up. Further doses of ¹³¹I RIT were given in patients who had persistent hyperthyroidism.

Definitions of Euthyroid, Hypothyroid, and Hyperthyroid Status

The patient’s laboratory report and their respective laboratory ranges were used to define the biochemical status. For patients with overlapping or conflicting laboratory results and those who could not be categorized definitely in one particular category, the subsequent follow-up information was used to classifying them.

Outcome Definitions

Outcome After First ¹³¹I Radioiodine Therapy

Cured (First-Dose Success)

Patients who achieve either a euthyroid state (which is stable for at least 1 year without the requirement of repeat RAI or antithyroid medication) or a permanent hypothyroid state requiring thyroxine supplementation after ¹³¹I therapy were defined as cured. Thus, combined euthyroid and hypothyroid patients were considered cured.

Not Cured (First-Dose Failure)

Persistent hyperthyroid after ¹³¹I therapy necessitating further treatment with radioiodine/antithyroid drugs.

Cumulative Cure Rates After Successive¹³¹I RIT

Cumulative frequencies of ¹³¹I RIT cure rates after successive RIT sessions (2nd, 3rd, 4th, or more RIT) were noted.

Estimation of Optimal First ¹³¹I Activity for the Treatment of High Turnover GD Patients

Estimating a Correction Factor for Commonly Used ¹³¹I Activity Calculation Formulae

Commonly used Marinelli or simplified Marinelli formulae for ¹³¹I activity calculations were considered [10]. Activity to be administered is inversely proportional to radioiodine uptake and effective half-life [Activity∝ 1/RAIU (max)*effective half-life].

Activity to be administered in the high turnover group can be adjusted by multiplying a correction factor X, where X is the ratio of T_eff for normal turnover GD and T_eff for the rapid turnover gland. To obtain this ratio, we need to estimate the effective half-lives for the normal turnover and high turnover groups.

Effective Half-Life (T_eff) Estimation

2-h RAIU and 24-h RAIU values were used to approximate the curve pattern. 2-h RAIU as the early time point seems more appropriate than 4–6 h, considering it is closer to the y-axis (as shown in supplementary figure 1, compared to 4 or 6 h) and shall provide a better approximation of uptake at time zero for curve fitting [30].

For effective half-life calculations, the exponential fitting method was used for curve fitting (an example is shown in supplemental figure 1). Since only two data points are used for curve fitting, only the subgroup of patients (55 patients) where the curve trend can be predicted (i.e., patients with 24-h RAIU less than 2-h RAIU showing a falling trend, were used). Patients with 24-h RAIU higher than or equal to or marginally less than 2 h were excluded (as T_eff calculation was not possible). Non-linear curve fitting (exponential fitting) was applied between two time points that give the slope (λ) value. From this, T_eff was obtained using formula T_eff = 0.693/λ, as shown in supplementary figure 1.

Statistical Analysis

Statistical analysis was performed using the Statistical Package for the Social Sciences for Windows (SPSS Version 26.0) (IBM, Chicago, IL, USA). The normalcy of data was checked using the Shapiro-Wilk W test. Mann-Whitney test was used to compare the non-normally distributed continuous variables between two groups. A chi-square test was used to compare the categorical variables between the two groups. Considering the effect of other covariates (age, gender, duration of disease, goitre severity, prior antithyroid drug use, ¹³¹I RAI activity) on the outcome proportions in two groups, matching covariates analysis was done using the propensity score method. We calculated the propensity score using logistic regression (turnover group as the dependent variable) and other covariates mentioned above as an independent variable to get the predictive probability. We then matched the patients by propensity score, using the nearest-neighbour method, with an equal ratio of 1:1, and the match tolerance taken was 0.02 [31, 32]. For checking the association of risk factors and the outcome, univariate and multivariate analysis was used. For univariate analysis (to check the association between individual explanatory variable and outcome variable in the absence of other explanatory variables) (both representing proportions), chi-square test was used, and odds ratio was calculated. For multivariate analysis (to find independence of an explanatory variable in the presence of other confounding variables), binary logistic regression method was used to find out the best fitting model to describe the relationship between an outcome (dependent variable) and a set of explanatory variables (independent variables). Receiver operating characteristic (ROC) analysis was used to find the cutoff value for continuous variables, using the Youden index (sensitivity + specificity-1) [33]. p value < 0.05 was considered significant. Time to event analysis, for the spontaneous hypothyroid conversion of euthyroid patients in the long-term follow-up after all RIT, using Kaplan-Meier estimates with the comparison between curves was done using log-rank statistics.

Results

A total of 312 patients recruited for this retrospective study: cohort 1 of 104 Graves’ disease patients with high turnover and cohort 2 as the control group (1:2 age/sex-matched normal turnover GD). The detailed flow chart is given in Fig. 1. The prevalence of high turnover GD in our population was 4%. The detailed characteristics of patients among the two groups are shown in Tables 1 and 2.

Table 1.

Baseline characteristics of patients in high turnover and normal turnover group included in the study

	High turnover	Normal turnover	p value
Number of patients	n = 104	n = 208
Age (years) (median, range)	38.5 (16–66)	38 (16–66)	0.916
Gender (male/female)	31(30%)/73 (70%)	66(32%)/142(68%)	0.729
Duration of disease (months) (median, range)	24 (1–192)	24 (1–150)	0.236
Prior antithyroid drug use	95(91.3%)	186(89%)	0.592
Duration of antithyroid drug use (months) (median, range)	24 (0–192)	18 (0–150)	0.312
Weight loss	96(92.3%)	183 (88%)	0.241
Palpitation	97(93%)	184(88.5%)	0.181
Increased sweating	81(78%)	168(81%)	0.550
Tremors	88(85%)	159(76%)	0.094
Diarrhoea	47(45%)	62(30%)	0.007
Proximal myopathy	11(10.6%)	13 (6.3%)	0.176
Thyroid eye disease	40(38.5%)	44(21%)	0.001
Proptosis grading
0	64 (62%)	164 (79%)	0.004
1	22 (21%)	28 (13%)
2	18 (17%)	16 (8%)
WHO goitre grading
0	2 (2%)	36(17%)	0.001
1	22(21%)	92 (44%)
2	80 (77%)	80 (39%)
First-dose RAI 1 activity (mCi) (median, range)	7 (5–15)	5 (5–15)	0.001
Radioiodine uptake
2-h RAIU (median, range)	62 (12–100)	25 (3–83)	0.001
24-h RAIU (median, range)	56 (11–100)	57.5(23–91)	0.029

Bold numbers imply significant p values

Table 2.

Baseline characteristics of patients after matching (matched to potential confounders using propensity score method)

	High turnover	Normal turnover	p value
Number of patients	n = 77	n = 77
Age (years) (median, range)	37 (16–66)	38 (17–64)	0.516
Gender (male/female)	21(27%)/56 (73%)	21(27%)/56(73%)	1.0
Duration of disease (months) (median, range)	24 (1–120)	24 (1–150)	0.928
Prior antithyroid drug use	69(90%)	67(87%)	0.616
Duration of antithyroid drug use (months) (median, range)	18 (0–120)	24 (0–150)	0.604
WHO goitre grading
0	2 (3%)	3(4%)	0.857
1	21(27%)	19 (25%)
2	54 (70%)	55 (71%)
First-dose RAI 1 activity (mCi) (median, range)	7 (5–15)	5 (5–15)	0.895

Two groups matched for potential confounders (age, sex, duration of disease, prior ATD use, goitre grading, RAI activity used)

Outcome After First RIT

The cure rate of 28% (16% hypothyroid and 12% euthyroid) in the high turnover cohort 1 compared to 66% (49% hypothyroid and 17% euthyroid) in the normal turnover cohort 2 was observed after the first RIT (p = 0.001). The detailed results are shown in Table 3. Matched pair analysis, based on the propensity score method, also showed similar results. The high turnover group’s cure rate was 31% compared to 61% in the normal turnover group (see Table 3).

Table 3.

Treatment outcomes, cure defined as hypothyroid or euthyroid at 12 months after first RIT

Age and sex-matched groups
High turnover (n = 104)		Normal turnover (n = 208)
Hypothyroid	17 (16%)	103 (49%)
Euthyroid	12 (12%)	35 (17%)
Cured*	29 (28%)	138 (66%) (p 0.001)
Complete multi-covariate# matched group (1:1) (after propensity score matching)
High turnover (n = 77)		Normal turnover (n = 77)
Hypothyroid	12 (15.5%)	34 (44%)
Euthyroid	12 (15.5%)	13 (17%)
Cured*	24 (31%)	47(61%) (p 0.001)

*Cure rate = hypothyroid rate + euthyroid rate

#Matched for age, sex, duration of disease, prior ATD use, goitre grading, RAI activity used

Cumulative Outcomes After Multiple RIT

In the high turnover cohort, patients required further RITs than the normal turnover group of patients. The median number of therapies, 2 RIT vs. 1 RIT, and median cumulated ¹³¹I activity (15 mCi vs. 7 mCi) were required in high turnover compared to normal turnover patients, as shown in Table 4.

Table 4.

Cumulative cure rates after multiple successive ¹³¹I radioiodine therapies

Cumulative success after multiple RIT	High turnover (102)*		Normal turnover (208)
	Cumulative frequency		Cumulative frequency
1 RIT	29 (28.4%)	28.4%	138(66.3%)	66.3%
2 RIT	44 (43.1%)	71.6%	54(26%)	92.3%
3 RIT	18 (17.6%)	89.2%	9 (4.3%)	96.6%
≥ 4 RIT	11 (11.1%)	100%	7 (3.4%)	100%
Median number of RIT required	2 (1–5)		1 (1–6)	(P 0.001)
Median cumulative 131I activity required (mCi)	15 (5–90)		7 (5–100)	(P 0.001)

*2 patients, after they received their last RIT (which is after 12th month) did not follow up, hence could not be assessed for the final response (cured/not cured), but for treatment outcome at 12 months, they were taken as a failure for First RIT

Among cured patients in the high turnover group, the final clinical outcome at the time of writing was 58% in hypothyroid and 42% in euthyroid metabolic states. In the normal turnover group, 71% were finally hypothyroid, and 29% were euthyroid, as shown in Fig. 1. There was a significant difference in the distribution of hypothyroid and euthyroid status in the two groups (p = 0.025).

The Fate of Euthyroid Patients After Completion of All RIT on Long-Term Follow-Up

Interestingly, in patients who were euthyroid after completing all RIT, with a median time duration of 65 months, 60% showed spontaneous conversion to hypothyroid state in the high turnover group. With a median time duration of 40 months, 69% in the normal turnover group become hypothyroid (p = 0.03), as shown in Fig. 2.

Fig. 2.

Time to event analysis (Kaplan Meier curves) (event: spontaneous conversion to hypothyroid state in euthyroid patients after completion of all RIT) showing the fate of euthyroid patients after completion of all RIT on long-term follow-up

Association of Pre-therapy Factors and Treatment Failure Rate After First RIT

Univariate and Multivariate Analysis

Continuous variables were categorized into categorical variables for modelling into univariate and multivariate analysis. We used the receiver operating characteristic (ROC) method to find the cutoff for the continuous variables used in the univariate and multivariate analysis model. The cutoff values for the age, 44 years; duration of disease, 11 months; duration of ATD use, 11 months; 2-h RAIU, 37%; and 24-RAIU, 57%, were used.

In univariate analysis, eight factors were found significant, namely, age > 44 years, duration of disease > 11 months, goitre grading, prior ATD use, high turnover, 2-h RAIU > 37%, 24-h RAIU > 57%, and amount of RAI activities were significantly associated with the first-dose failure, while in multivariate analysis, only 4 factors were independently associated with first-dose failure, namely, the age > 44 years (odds ratio 2.6, p = 0.001), grade 2 goitre (odds ratio 2.8, p = 0.029), high turnover group (odds ratio 2.6, p = 0.007), and 2-h RAIU > 37% (odds ratio 2.7, p = 0.005); the details are shown in Table 5.

Table 5.

Univariate and multivariate analysis of the association between risk factors and treatment failure after first RIT

Variables	First RAI outcome		Univariate analysis		Multivariate analysis
	Not cured	Cured	Odds ratio for failure (CI)	(p value)	Adjusted odds ratio (CI)	(p value)
Age
≤ 44 years > 44 years	80 (41%) 65 (56%)	117 (59%) 50 (44%)	1 1.9 (1.1–3.0)	0.007	1 2.6 (1.4–4.5)	0.001
Gender
Female Male	100 (46%) 45 (46%)	115 (54%) 52 (54%)	1 0.9 (0.6–1.6)	0.984	–
Duration disease
≤ 11 months > 11 months	26 (35%) 119 (50%)	48 (65%) 119 (50%)	1 1.8 (1.07–3.1)	0.025	1 1.3 (0.6–2.5)	0.483
Goitre
0	8 (21%)	30 (79%)			1
1 2	37(32%) 100 (62%)	77 (68%) 60 (38%)	n.a	0.001	1.1 (0.4–2.9) 2.8 (1.1–7.4)	0.782 0.029
Prior ATD
No Yes	8 (26%) 137 (49%)	23 (74%) 144 (51%)	1 2.7 (1.1–6.3)	0.015	1 2.6 (0.9–7.4)	0.06
Turnover
Normal High	70 (34%) 75 (72%)	138 (66%) 29 (28%)	1 5 (3–8.5)	0.001	1 2.6 (1.2–5.2)	0.007
RAIU (2)
≤ 37 > 37	43 (27%) 102 (68%)	118 (73%) 49 (32%)	1 5.7 (3.5–9.3)	0.001	1 2.7 (1.3–5.5)	0.005
RAIU (24)
≤ 57 > 57	62 (39%) 83 (55%)	98 (61%) 69 (45%)	1 1.9 (1.2–2.9)	0.005	1 1.2(0.6–2.2)	0.524
RAI 1 (mCi)
5–7	103 (43%)	138 (57%)			1
8–10 11–15	28 (55%) 14 (70%)	23 (45%) 6 (30%)	n.a	0.027	0.7 (0.3–1.4) 0.7 (0.2–2.2)	0.334 0.390

RAI radioactive iodine, OR odds ratio, CI confidence interval, ATD antithyroid drug, RAIU radioactive iodine uptake, mCi milli Curie. Calculation of odds ratio is possible in dichotomous variables; for goitre grading and RAI 1 activity variables, only significance value is mentioned; odds ratio is not mentioned

The median cumulated ¹³¹I activities required for the categorized groups based on these significant variables are shown in Table 7, which shows that the highest activity was needed for the high turnover group.

Table 7.

Median cumulated ¹³¹I activity required for cure in categorical groups of independently significant variables

Independently significant variables in multivariate analysis	Significance (p value)	Median cumulated 131I activity ultimately required for cure in categorized groups
Age
≤ 44 years > 44 years	0.001	7 mCi 10 mCi
Goitre
0 1 2	0.029	5 mCi 7 mCi 12.5 mCi
Turnover
Normal High	0.007	7 mCi 15 mCi
RAIU (2)
≤ 37 > 37	0.005	5 mCi 14 mCi

For the high turnover group, the adjusted odds ratio in multivariate analysis (multivariate-adjusted) was 2.6 (95% CI: 1.2–5.2), p = 0.007, as shown in Table 5. This finding was also confirmed using another method using propensity score-matched data (PSM adjusted odds ratio was 3.5 (95% CI: 1.7–8.2, p = 0.001) (details shown in Table 6).

Table 6.

Analysis to find association between risk factors and treatment failure after first RIT in propensity score matched data

Variables	First RAI outcome		Univariate analysis on propensity score matched
	Not cured	Cured	(PSM) data PSM adjusted odds ratio for failure (CI) (p value)
Age
≤ 44 years > 44 years	48 (48%) 35 (65%)	52 (52%) 19 (35%)	1 1.9 (1.008–3.9)	0.046
Gender
Female Male	62 (55%) 21 (50%)	50 (45%) 21 (50%)	1 0.8 (0.4–1.6)	0.553
Duration disease#
≤ 11 months > 11 months	17 (45%) 66 (57%)	21 (55%) 50 (43%)	1 1.6 (0.8–3.4)	0.192
Goitre
0	8 (21%)	30 (79%)
1 2	37(32%) 100 (62%)	77 (68%) 60 (38%)	n.a	0.01
Prior ATD
No yes	6 (33%) 77 (57%)	12 (67%) 59 (43%)	1 2.6 (0.9–7.3)	0.06
Turnover
Normal High	30 (39%) 53 (69%)	47 (61%) 24 (31%)	1 3.5 (1.7–8.2)	0.001
RAIU(2)
≤ 37 > 37	15 (27%) 68 (70%)	41 (73%) 30 (30%)	1 6 (3–12)	0.001
RAIU (24)
≤ 57 > 57	33 (50%) 50 (57%)	33 (50%) 38 (43%)	1 1.3 (0.7–2.5)	0.41
RAI 1 (mCi)
5–7	63 (54%)	54 (46%)
8–10 11–15	14 (50%) 6 (67%)	14 (50%) 3 (33%)	n.a	0.683

PSM propensity score matching, CI confidence interval

Prediction of Optimal First Dose of RIT to Be Used in High Turnover Group

The mean effective half-life in the high turnover group obtained by the exponential fitting method was 3.5 days (standard deviation: +1.79 days). This T_eff was consistent with the values mentioned in the literature, which is the effective half-life of 2.8 days in the high turnover gland and around 5.5 days in the normal turnover gland [14–16].

Estimation of the Correction Factor for Dose Calculation in High Turnover GD

The correction factor was defined as the ratio of T_eff for normal turnover GD to T_eff for the rapid turnover gland. Using, T_eff for normal turnover GD as 5.5 days, as taken from the literature [14–16], we get a multiplication factor of 1.5 (when T_eff for high turnover taken as 3.5 days, which we got in our study) and 1.96 (when T_eff for high turnover taken as 2.8 days which is mentioned in published literature) [14–16]. Hence, the correction factor is around 1.5–2.

Discussion

A high turnover of iodine is clinically significant because the decreased retention time of ¹³¹I not only reduces the radiation dose delivered to the thyroid gland (target organ) that results in therapeutic failure but also results in increased radiation dose to non-target (blood and bone marrow) tissues [12, 34, 35].

It has been noted that up to 12–20% of GD patients may have a rapid ¹³¹I turnover in the thyroid gland [13, 36, 37]. In our study, we got a rapid turnover prevalence rate of around 4% of cases of GD, which is significantly less as compared to the values mentioned above in the literature. The probable reason could be due to the different definition of rapid turnover based on early and late RAIU ratio used in various studies. In our research, we have used 2 h as early RAIU, while in the study by Aktay et al., authors have taken 5 h as early RAIU [13]. RAIU measurement strongly depends on iodine levels that vary with the environmental iodine content of a geographic region’s food and drinks which can produce disparity in RAIU values [25, 38, 39].

Studies evaluating the effects of prior ATD have shown conflicting results, with one group of investigators showing negative impact [14] and others leading no effect [40, 41]. In our study, prior ATD use was not significant in multivariate analysis; however, it significantly impacted RIT results in normal turnover patients when analysed separately in each cohort, but not in high turnover patients.

Few prior studies have suggested that in patients with young onset of GD, there is an increased likelihood of relapse after ATD treatment [42, 43]. Interestingly, in our study, we observed that older age (> 44 years) patients had 2.6 times more odds of failure to first RIT than younger age (18–44 years) patients.

Zhang et al., in their study, showed a goitre size of more than 56 g which is associated with a high turnover of the gland [30]. In their research, DeGroot et al. also concluded that larger thyroid volumes need to be treated with higher dosages of ¹³¹I per unit volume [44], and this has been confirmed by other studies too [45]. In our previous study, Damle et al. [46], studied the effect of RAIU on ¹³¹I therapy outcomes where the authors administered 5 mCi ¹³¹I (mean activity) for the treatment of GD and reported the first-dose success rate of 81.7% in patients with 24-h RAIU < 50% and 68.6% in the group with 24-h RAIU >  50% (p <  0.001).

RAIU is an important predictive factor for determining treatment outcome. Only a single RAIU value does not give information about the thyroid gland’s iodine kinetics, only give an idea about iodine avidity. Rather, the ratio of early to late RAIU can give an idea about the gland’s turnover. Initially, it was believed that maximum uptake of RAI occurs at 24 h, and few studies evaluated the impact of a single RAIU (24 h) on RIT outcome. In our previous study, as reported by Damle et al., the effect of RAIU on ¹³¹I therapy outcomes [46]. The authors administered 5 mCi ¹³¹I (mean activity) for the treatment of GD and reported the first-dose success rate of 81.7% in patients with 24-h RAIU < 50% and 68.6% in the group with 24-h RAIU > 50% (p < 0.001). Similar results were shown in another study by Kristoffersen et al. [47]. In our research, though the significant association between 24-h RAIU and RIT failure was found in univariate analysis (p 0.005), on multivariate analysis, it was not significant (p 0.52), rather only 2-h RAIU was significantly associated with treatment failure in multivariate analysis {OR-2.7 (95% CI: 1.3–5.5), p = 0.005}. Both these studies mentioned above showed the importance of RAIU in estimating RIT administered activity, but the importance of early RAIU was not evaluated previously; this is the first study highlighting the value of early 2-h RAIU in Graves’ disease patients. However, a high ¹³¹I turnover has been demonstrated as a strong predictor of RIT failure in several studies [13, 48, 49].

We found significantly decreased success rates after the first RIT in the high turnover group (28%) compared to the normal turnover group (66%) (p = 0.001). Aktay et al., in their study, found that up to 50% of the GD patients with high¹³¹I turnover failed to respond to the initial ¹³¹I therapy, hence required multiple RAI therapies [13].

The amount of administered ¹³¹I activity plays a significant role in GD treatment success in the high turnover group, as shown in Table 7. In this context, Aktay reported that a mean dose of 28 mCi was required to treat patients with high turnover [13]. A recent multi-arm RCT that compared low dose RIT (3.7 MBq/g) with lithium, moderate dose ¹³¹I (5.55 MBq/g) with lithium, against the high dose RIT (7.4 MBq/g) without lithium. Among the groups, patients belonging to the third group had the best outcome where the median dose of ¹³¹I was 25 mCi [20].. The authors achieved cure rates of 80% with the median use of 25 mCi (12–49 mCi). In our study, the median cumulative activity of 15 mCi was required in the high turnover group compared to the median 7 mCi in the normal turnover group to achieve a cure.

Higher ¹³¹I activity results in higher cure rates but shall result in a more hypothyroid state [18, 50]. Achieving euthyroidism following RIT is the best outcome; however, it is challenging to predict whether patients shall achieve euthyroid or hypothyroid metabolic state [51, 52]. We too observed that among all cured patients in our series, more patients were hypothyroid in the normal turnover group (71%) than in the high turnover group (58%) after completion of all RITs; another interesting fact is that the patients who were euthyroid after completion of all RIT, when followed for longer duration, had higher hypothyroid conversion rate and also took shorter duration to achieve that feat (median time 40 months versus 65 months, respectively). These observations could be explained by the fact that rapid turnover kinetics does not allow radioiodine to stay in the gland for a longer duration to deliver higher radiation absorbed dose to the thyroid gland; thus, actually lesser absorbed dose is delivered. However, in the normal turnover gland, a higher absorbed dose is delivered (mCi-to-mCi basis) compared to the rapid turnover gland, resulting in quick cure and more hypothyroidism for similar administered ¹³¹I activity keeping all other variables identical.

No precise dose calculation can secure RIT’s dual purpose in Graves’ disease, i.e. achieving a high first-dose ¹³¹I success rate and long-term euthyroidism; hence, it is found acceptable to offer a fixed quantity of RAI. The administered RAI should be corrected for high turnover because these patients frequently show a failure to activities routinely administered for normal turnover patients. The administered activity can be done by multiplying a correction factor if high turnover is present. The correction factor of 1.5–2 should be multiplied to fixed-dose activity calculations for high turnover GD. This proposed correction factor derived from the effective half-life ratio and activity calculation formulas correlated well with the results of actual treatment and outcomes results seen in our study. In our research to achieve a cure in 90% of patients, the median cumulative activity required was 15 mCi in the high turnover group, approximately two times of 7 mCi required in the normal turnover group. The contribution of each variable is depicted in Table 7.

The highest cumulated activity was required for high turnover group patients, among other independently significant variables. Using a correction factor to adjust for high turnover will take care of other variables while planning to administer ¹³¹I activity.

Strength of the Study

The strength of our study is that a large number of patients to compare two groups (high turnover cases and normal turnover controls) were included, assuming the power of study at 95% and type 1 error at 5%, making it a robust study from a statistical point of view. The effects of covariates were studied and confirmed using two robust statistical methods: (a) using a logistic regression model in heterogeneous groups/cohorts’ original data and (b) baseline matched cohorts using propensity score matching to get uniform groups.

Limitations

The study is not without limitations; first, the study design is a retrospective cohort study. There was a lack of data on anti-thyrotrophin receptor antibody (TRAb) status. For effective half-life calculations, only two time points, 2 h and 24 h, were used. The patients’ exact iodine status, whether the patients were iodine sufficient or iodine-deficient at the time of treatment, was not assessed.

Conclusion

High turnover Graves’ disease differs significantly from a normal turnover counterpart in post ¹³¹I outcomes. Factors associated with high failure rate are higher age (> 44 years), higher grade of goitre, higher 2-h RAIU (> 37%), and high turnover of gland. Based on the gland’s turnover, first administered ¹³¹I activity can be increased by 1.5 to 2 times in the high turnover group to improve the ¹³¹I outcome, which adjusts for other variables.

Authors’ Contributions

Both authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Saurabh Arora and Chandrasekhar Bal. The first draft of the manuscript was written by Saurabh Arora, and Chandrasekhar Bal commented on previous versions of the manuscript. Both authors read and approved the final manuscript.

Data Availability

Original raw data is available on request to corresponding author.

Compliance with Ethical Standards

Conflict of Interest

Saurabh Arora and Chandrasekhar Bal declare that they have no conflict of interest.

Ethical Approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the Helsinki declaration as revised in 2013 and its later amendments or comparable ethical standards. For this type of study, formal consent is not required.

Informed Consent

The institute ethics committee (Human studies), All India institute of Medical Sciences, New Delhi, approved this retrospective study (ref no: IECPG-598/24.10.2019) and the requirement to obtain informed consent was waived.

Consent to Participate

Not applicable.

Consent to Publish

Not applicable.