- Top of page
- PATIENTS AND METHODS
Desmopressin, a synthetic analogue of the antidiuretic hormone vasopressin, has been used for several years in the treatment of diabetes insipidus and nocturnal enuresis. The safety profile of desmopressin is well known. The only potentially serious adverse effect is hyponatraemia. Desmopressin alone does not induce hyponatraemia but allows it, as water continues to be conserved even if the patient's fluid intake exceeds physiological needs. Recently, the use of desmopressin for nocturia (waking to void at night) was investigated in a large scale study . In clinical practice most patients with nocturia are likely to be elderly, because the prevalence of nocturia increases with age [2–11]. The safety profile established in children and adults cannot be directly transferred to the elderly as, in general, they have more concomitant diseases and use more drugs than do younger subjects. With age there are changes in the regulation of water and electrolyte balance, making the elderly more vulnerable to drug- and/or disease-induced hyponatraemia . The primary aim of the present study was to investigate the short-term safety of desmopressin in elderly patients with nocturia, focusing on risk factors for hyponatraemia. The effects of desmopressin on urine output, sleep and voiding patterns were also investigated.
PATIENTS AND METHODS
- Top of page
- PATIENTS AND METHODS
Eligible responders from a questionnaire study on nocturia in people aged ≥ 65 years and living in Tierp, Sweden, were enrolled in a study using frequency-volume charts (FVCs) . Among patients who submitted completed FVCs, all those with nocturia (defined as two or more voids per night) were contacted by telephone and asked about participating in a treatment trial with desmopressin. Written informed consent was obtained before any study procedures were commenced. Screening included a medical history, concomitant medication, physical examination, measuring supine and standing blood pressure, blood electrolytes and urine analysis (including spot urine osmolality). Patients with a serum sodium level below the normal range, epilepsy, clinically significant renal or hepatic disease (as evidenced by dialysis or ascites) were excluded. The patients were instructed to take one 0.2 mg desmopressin tablet at bedtime for three consecutive nights and to drink enough only to satisfy their thirst from 1 h before until 8 h after taking the tablet. They were also given a FVC and instructed to record bedtime, waking time, and the time and volume of each void for 72 h, starting after taking the first tablet. Serum sodium, serum osmolality and body weight were measured before noon on the day after the first and third dose. Adverse events were collected by spontaneous report and with the answer to the question: ‘Have you had any health problems since your previous visit?’. One patient undertook the study twice because of a suspected laboratory failure during the first attempt. For all variables recorded twice for this patient the mean of the two measurements was used in the analysis.
The 3-day FVC obtained from the previous study  was used as the baseline for all diary-derived variables. Covariates and potential predictors of hyponatraemia, and pharmacological and clinical response were assessed using multiple and logistic regression. As true hyponatraemia may not have time to develop in a 3-day trial the proxy used was ‘sensitivity to change in serum sodium’, defined as a mean serum sodium level during treatment deviating by more than five units from baseline, and this was used in the logistic model. The pharmacological response was measured as change from baseline in nocturnal diuresis, defined as the nocturnal urine volume divided by night length. The clinical response was measured both as the change from baseline in the number of nocturnal voids and as the change from baseline in the time from bedtime to the first nocturnal void (or waking, if no nocturnal void occurred). Possible explanatory variables were screened using linear regression for continuous response variables and logistic regression for binary response variables. The inclusion criterion for further modelling was P < 0.15. The final model was derived using stepwise selection amongst the variables included above, with a selection criterion of P < 0.10. Variables remained in the final model if their corresponding P value was < 0.10 . All variables screened for and included in the stepwise selection procedure are presented in Table 1.
Table 1. Demographic, baseline and other characteristics screened for (scr) and included in (incl, i.e. P < 0.15 in screening) the stepwise selection logistic and multiple regressions
|Body mass index||incl||incl||scr||scr|
|Urine leukocytes (≥ +1)||scr||scr||incl||scr|
|Baseline|| || || || |
|24-h urine output||incl||incl||incl||incl|
|Largest voided volume||scr||incl||incl||scr|
|No. of nocturnal voids|| – || – ||scr|| – |
|Time to 1st noct void|| – || – || – ||scr|
|% nocturnal/24-h urine||scr||incl||scr||scr|
|24-h urine output||scr||scr||scr||scr|
- Top of page
- PATIENTS AND METHODS
Of the 599 people who reported nocturia in the questionnaire study, 271 were mailed an FVC; among the 159 who submitted their FVC, 116 had two or more voids per night, and 73 were screened and 72 enrolled. All enrolled patients completed the study. One patient was excluded from the efficacy analysis because of a faulty baseline diary. The study population was representative of the previously studied nocturia population  in terms of voiding and sleep characteristics, but not in gender distribution, as there was a higher proportion of men.
The demographic and baseline sleep and voiding characteristics are summarized in Table 2. The study population comprised 53 men (74%) and 19 women (26%), with a median (range) age of 75.5 (66–90) years; all were Caucasian. The reported alcohol consumption was low, at a mean (sd) of 0.8 (1.5) units/week (maximum eight) and smoking was only reported by one patient. All patients were ambulant and lived at home. Concomitant diseases and medications in the study population are outlined in Table 3. The physical examination showed one previously undiagnosed case of congestive heart failure and one of Parkinson's disease. Only 14 patients had a completely blank urine dipstick test; bacteriuria was found in eight patients (11%), but 26 (36%) had signs of leukocytes in their urine.
Table 2. Demographic and baseline voiding characteristics of the study population
|Height, cm||72|| 171 (8)|
|Body weight, kg||72|| 79.2 (12.2)|
|Body mass index, kg/m2||72|| 27.2 (4.2)|
|No. of 24 h voids||70|| 8.8 (1.7)|
|No. of nocturnal voids||72|| 2.7 (0.7)|
|24-h urine volume, mL||69||1637 (427)|
|24-h urine volume/weight, mL/kg||69|| 21.6 (5.9)|
|Nocturnal urine volume, mL||71|| 748 (263)|
|Nocturnal urine volume/weight, mL/kg||71|| 9.8 (3.4)|
|Nocturnal diuresis, mL/min||71|| 1.4 (0.5)|
|Largest voided volume, mL||71|| 339 (102)|
|Mean voided volume, mL||71|| 192 (56)|
|Duration of night, h||72|| 9.0 (0.8)|
|Time to 1st nocturnal void, h||72|| 2.7 (0.9)|
Table 3. Selected concomitant diseases and medications in the study population
|Cardiovascular diagnosis*||25 (35)|
|Diabetes mellitus|| 7 (10)|
|Urological diagnosis†||25 (35)|
|Orthostatic hypotension‡||32 (44)|
|Bacteriuria|| 8 (11)|
|Low dose acetylsalicylic acid||14 (19)|
During the study, 34 patients (47%) reported a total of 49 adverse events; headache was most frequent (16% of patients) and was in all cases rated as mild. No serious adverse events occurred and none of severe intensity were recorded. The mean serum sodium level decreased during treatment (Fig. 1), the decrease being most pronounced after the first dose. A serum sodium below the normal range was registered in four patients (Table 4).
Figure 1. The serum sodium values at baseline and after the first and third dose of desmopressin. The central line is the median, the box the interquartile range, the bars the 10th and 90th percentile and the points the outliers.
Download figure to PowerPoint
Table 4. Details of patients classified as sensitive to change in serum sodium or with serum sodium below the normal range (134–146 mmol/L)
|Sex/age, years||Serum sodium (mmol/L) |
|Baseline||After 1st dose||After 3rd dose|
|Sensitive to change in serum sodium|
|Serum sodium below normal range|
After the first dose, one patient had a decrease in serum sodium from 141 to 130 mmol/L. There were no significant changes in serum osmolality or weight and no symptoms except a mild headache were reported. Desmopressin was continued and after the third dose the serum sodium was 145 mmol/L. The low value was suspected to be a laboratory failure, and this patient was included in the study again and all study procedures repeated. The decrease in serum sodium was not reproduced.
Six patients qualified as being sensitive to change in serum sodium (Table 4). The model obtained with logistic regression for sensitivity to change in serum sodium had a good fit and predictive power (c = 0.90) . The risk (odds ratio, 95% CI) increased with increasing age (1.3, 1.1–1.6), concomitant cardiac disease (10.0, 0.9–105.8) and increasing baseline 24-h urine output (1.2, 1.0–1.5).
All six patients classified as sensitive to serum sodium were aged ≥ 77 years and four of them had a concomitant cardiac diagnosis (angina pectoris, atrial fibrillation and previous myocardial infarction with and without angina pectoris). One patient in the study had a baseline 24-h urine output of> 40 mL/kg body weight and was deemed sensitive; all other sensitive patients had a 24-h urine output of < 30 mL/kg.
The mean 24-h urine output decreased significantly more in those sensitive to change in serum sodium, with a mean (sd, 95% CI) change in 24-h urine output of 693 (330, −1040 to −347) mL, than in the insensitive group, at − 180 (289, − 250 to −110) mL. All patients sensitive to change in serum sodium had a reduction of> 40% in nocturnal urine output during treatment.
The efficacy results are summarized in Table 5; the mean number of nocturnal voids decreased during treatment but no patient had complete relief of nocturia (i.e. no nocturnal voids). Nocturnal urine output decreased while the time to first nocturnal void increased during desmopressin treatment. There was a reduction of ≥ 20% in nocturnal diuresis in 58 patients (82%) and of ≥ 50% in 31 (44%). There was an increase in time to first nocturnal void of ≥ 50% in 38 patients (54%).
Table 5. The change from baseline in sleep and voiding variables
|Variable||Mean (sd, 95% CI difference)|
|Nocturnal voids, n|| − 1.0 (0.6, − 1.1 to – 0.9)|
|Nocturnal urine volume, mL|| − 306 (193, − 351 to −261)|
| per body weight, mL/kg|| − 4.2 (2.5, − 4.8 to −3.6)|
| per time, mL/min|| − 0.5 (0.4, − 0.6 to −0.4)|
|Time to 1st nocturnal void, h|| 1.9 (1.8, 1.5–2.3)|
Response measured as the change in nocturnal diuresis increased with increasing baseline nocturnal diuresis and decreased with increasing nocturia index, baseline 24-h volume and urine osmolality. However, the model fit was poor (R2 = 0.67). The multiple regression models obtained for the change in nocturnal voids and in time to first nocturnal void were not evaluable as the fit was poor (R2 = 0.19 and 0.39).
- Top of page
- PATIENTS AND METHODS
Desmopressin was well tolerated in this short-term trial; the risk of hyponatraemia, as measured by the proxy, increased with age, cardiac diagnosis and higher baseline 24-h urine output. The results for age accord with previous research showing an increased general risk of electrolyte disturbances in the elderly . Desmopressin reduced nocturnal diuresis and the number of nocturnal voids, and increased the time from bedtime to the first nocturnal void. The results for nocturnal diuresis and voids were of the same magnitude as in previous findings, while the change in time from bedtime to first nocturnal void was greater [1,16].
The primary aim of the present study was to investigate safety in an unselected nocturia population, such as could be seen in a general practice, and exclusions were therefore minimised. Previous trials in this area have mostly focused on small segments of the potentially treatable population, often excluding those with common diseases or using frequently prescribed medications. During recruitment it soon became evident that men were substantially more willing to participate in the trial than were women. This was reflected in the gender composition of the group and further trials containing more women are needed to assess any gender differences in safety.
Hyponatraemia is defined as a deficiency of sodium in the blood; desmopressin as such does not produce hyponatremia but allows it, as it blocks the mechanism of compensatory diuresis. As long as fluid intake balances fluid loss, continuous desmopressin therapy can be used safely, as confirmed by its use in patients with diabetes insipidus. However, if fluid intake is not properly curtailed, dilutional hyponatraemia may result. Once-daily dosing is supposed to provide an antidiuretic effect only during the night, thus allowing for compensatory diuresis in the day even if fluid intake is excessive. The only precautionary measure taken in this study to avoid hyponatraemia was the recommendation to drink only when thirsty during the night.
To be successful, the search for risk factors for hyponatraemia requires that the correct variables are measured. It was anticipated that the rate of actual clinically significant hyponatraemia after 3 days would be too low to be useful in finding risk factors. Therefore the proxy (sensitivity to change in serum sodium) was chosen in the protocol, even though no data for its correlation with actual clinically significant hyponatraemia were available. The serum sodium value after one dose in the present trial was a poor predictor of sensitivity. Fifteen patients had an initial decrease of ≥ 4 units, among whom were the six patients classified as sensitive to change in serum sodium. Of these 15 patients, seven continued to have a decline, two stabilised and six had an increase in their sodium level after the third dose. Additional analysis with multiple regression of the change in serum sodium from baseline to the morning after the third dose gave no results, because the fit was poor. The logistic regression on sensitivity had a good model fit. The dominating risk factor was age; the youngest sensitive patient was 77 years old. In all, 28 patients were aged ≥ 77 years, of whom six were classified as sensitive. Thus it is probable that the key factor is not age per se but some underlying factor(s) increasing in prevalence with age. Neither orthostatic hypotension, hypertension, use of diuretics or NSAIDs showed any correlation with sensitivity (Table 1). The presence of a cardiac diagnosis and a high 24-h urine output were both incorporated, but neither replaced age in the model. Patients sensitive to change in serum sodium were pharmacological responders and had a larger decrease in 24-h urine output than the insensitive patients, indicating that the drug effect was not limited only to the night. Theoretically, a 24-h suppression of diuresis is required to attain dilutional hyponatraemia during desmopressin treatment. Whether this failure to achieve compensatory diuresis in the day is the result of a prolonged duration of desmopressin or endogenous antidiuresis in the day remains to be investigated.
Caution is warranted in interpreting the efficacy results of this study as it was not optimized for investigating efficacy. There was no placebo arm, and the baseline and treatment assessments were separated in time. However, the results are comparable with those in other trials of desmopressin for this indication. A specific source of efficacy bias concerns the safety instructions given in connection with desmopressin treatment. Patients were instructed to drink only to satisfy thirst, from 1 h before until 8 h after taking desmopressin; these instructions could affect fluid intake and thus also urine volume and the number of nocturnal voids. None of the patients reported having any problem with the instruction and many stated that they had already curtailed fluid intake in the evening as a way to avoid excess nocturia. The habit of voluntary fluid restriction has previously been described in patients with nocturia  and incontinence . The failure to find any strong connection between the response and the variables assessed is perhaps not surprising. Desmopressin, a vasopressin analogue, induces re-absorption of water in the collecting ducts of the kidney, which increases urine osmolality and decreases urine volume. Thus desmopressin would not be expected to significantly improve nocturia primarily caused by a small functional bladder volume. Neither could a high overnight urine volume primarily caused by solute diuresis, nor by insensitivity to vasopressin in the kidneys, be expected to respond significantly. Spot urine osmolality had some correlation with pharmacological response and further research on the predictive value of overnight urine osmolality could be of interest.
The use of desmopressin for nocturia is increasing, because of published reports and the registration of desmopressin for nocturia in several countries. The increased sensitivity to change in serum sodium with increasing age noted in the present trial indicates a need for close surveillance of elderly patients, especially if there is a concomitant cardiac diagnosis and/or a high 24-h urine output. Age per se is not thought to be the key factor in hyponatraemia and it is likely that the main underlying factors remain to be found. However, the strong association with age calls for a pragmatic approach, i.e. serum sodium testing in the elderly before and after a few days on treatment.