PREVENTING URINARY CATHETER BLOCKAGES Part 1 of 2

BARRY SIMPSON

According to the users' information leaflet, Nitrofurantoin (also known as Macrobid, Macrodantin or Genfura) is used 1) to cure urinary infections and 2) to prevent them. It is proved below that if used to prevent urinary infections, in my case it also prevented catheter blockages. I do not know for how many other people with catheter blockages it would be effective. It would be unlikely to work for those who have blockages caused by kidney stones or bladder stones.

The normal dose for preventing urinary infections is 50 or 100mg daily at night. Having taken 50mg per day for 30 days, this was reduced to 50mg 3 or 4 times per week and then to just nights when I was feeling feverish as might warn of a urinary infection or when there was a considerable amount of sediment in my catheter. That worked for me, but others might need different doses. Like other medications, Nitrofurantoin comes with many cautions and possible side-effects, listed in the users' information leaflet, but I have not had any at these doses.

As well as taking Nitrofurantoin I also repositioned my catheter by pulling it forward immediately after going to bed to prevent the intake being obstructed by pressing up against my bladder wall and, as far as possible, assisted drainage by gravity by placing my catheter flat on the bed rather than strapped to my leg.

I have also used several supporting methods to keep my catheter clear. These are listed below.

WHAT CAUSED THE BLOCKAGES?

From January to July 2016 the misery of my spinal injury was aggravated by the torment of catheter blockages. Here are a few observations to help identify what had been causing them:

1 My suprapubic catheter was installed in May 2013, about 4 months after my spinal injury. I had no blockage during the first two years and eight months but I had 23 between 17/1/16 and 24/7/16. Towards the end of that period, they became more frequent.

2 All the blockages occurred soon after changing positions from sitting upright in my wheelchair to lying flat on my back in bed. On every occasion I have wakened up sweating and trembling with exceptionally violent spasms, usually between midnight and 1am. There is a highly significant relationship between time and occurrence of blockages. If we use simple dichotomy that blockages could occur at either day or night, if there were no relationship between time of blockage and occurrence, the probability of any one blockage occurring at night would be 0.5. The probability of all 23 blockages occurring at night would be 0.5 to the power 23 = 0.000000119 or a little over 1 chance in ten million.

http://www.rapidtables.com/calc/math/Exponent_Calculator.htmThe base is 0.5 and the exponent is 23.

The binomial distribution can also be used with the same result:

http://www.vassarstats.net/textbook/ch5apx.htmlwhere N = 23, k = 23, p =0.5 and the answer is p(k out of N)

The multinomial can be used too:

https://www.easycalculation.com/statistics/multinomial-distribution.phpwhere the number of outcomes is 2 (night/day), the number of occurrences 23 and 0, p = 0.5 for each outcome.

Sitting is a wheelchair, my catheter is usually strapped to the upper side of my left leg. If left there when lying flat in bed, this part of my catheter might be at a higher level than the intake, impeding flow by gravity. I have noticed that sweating has on a few occasions been stopped by unstrapping my catheter from my leg and laying it at a lower level on the bed. Now I always have it unstrapped when in bed.

3 The nurse has never taken more than a few minutes to unblock the obstruction. It takes about 10 minutes for the sweating and spasms to subside enough to be clearly noticeable and a further 10 minutes for the sweating to have gone and the spasms returned to their normal level. In perhaps 8 or 10 of my 23 cases of blockage, it was uncertain why the blockage had cleared. Sometimes there was insufficient sediment for that to be a likely cause. Movement of the catheter in some cases seemed to be sufficient to make the urine to flow. Unblocking the catheter did not in every case cause a sudden, clearly visible flow of urine.

4 Only once has a blockage recurred during the same night. On this occasion, two nurses arrived at 10.50pm and replaced my catheter which was choked with 'sludge'. At 2.40am I woke sweating again. The same nurses returned. This time the blockage was 'positional'.

5 There were at least two causes of the blockages - bacterial and physical obstruction.

BACTERIAL BLOCKAGES

6 For about half of the blockages, the nurse mentioned sediment and/or 'pus' as the likely cause. On some occasions, including the most recent, only 36 hours after a bladder wash-out, the nurse said there was no sediment.

7 My leg bag is supposed to be changed weekly. On 3 or 4 occasions, the nurse has mentioned a dirty leg bag, when the carers (and me) had forgotten about it or when we had run out of them.

8 On the morning following several of the blockages I have had a bout of sweating which has always been stopped by taking Nitrofurantoin. This seems to indicate that I had a urinary infection, and that the infection was caused by bacteria which also caused the blockage the night before. A urine sample has never been taken at the time of a blockage or soon after. However, it seems likely that the bacteria killed by Nitrofurantoin created the sediment which resulted in most of the blockages.

I do not have a good explanation as to why I had no blockages between installation of my catheter in May 2013 and 17/1/16 despite having more than a dozen urinary infections during this period. I can only state the obvious that whatever bacteria were causing these infections did not produce sufficient sediment to cause a blockage but there was a change in January 2016. Before July 2016 I took Nitrofurantoin only occasionally to cure urinary infections.

When Nitrofurantoin stopped my blockages it also removed the additional spasms accompanying the infection and the blockage. So Nitrofurantoin achieve four results for me: i) it cured urinary infections; ii) it prevented urinary infections; iii) it prevented blockages; iv) it cured/prevented additional spasms resulting from infections and blockages.

BLOCKAGES BY PHYSICAL OBSTRUCTION

9 Weekly bladder wash-outs were started soon after the blockages began. I have never had a blockage on the following night, but I have had them soon after.

10 For about half the blockages, the nurse mentioned a physical obstruction, such as a collapsed tube. I do not know why physical obstructions were absent from May 2013 until January 2016. Maybe there was a change in type or brand of catheter in December 2015 or January 2016?

11 For the most recent blockage, the nurse offered the following explanation: the end of the catheter tube might be pressing up against my bladder wall, so obstructing the intake. This fits well with what two or three other nurses have said: 'I'm not sure what caused the blockage but wiggling the catheter where it enters the bladder seemed to unblock it'.

AN INITIAL TRIAL: 15th JULY - 12th AUGUST 2016

From 17/1/16 to 24/7/16 I had 23 blockages, including 6 from 15 - 24/7/16. It seemed that infections by bacteria were responsible for the majority of the blockages (point 8 above) and the position of my catheter for the others (point 11). It is likely that some blockages were caused by a combination of the two: a constriction not severe enough to block the flow of urine caused sediment to collect on the upstream side and this caused a blockage.

So it is very likely that changing positions from wheelchair to bed triggered nearly all the blockages, most of which were caused by sediment in the catheter. Why might changing positions cause an accumulation of sediment in my catheter? Gravity would probably act more weakly in bed; sediment in my bladder might change position on going to bed and obstruct the intake of my catheter. These are possible explanations as to why changing positions from wheelchair to lying flat in bed seemed to trigger blockages but I do not have a convincing answer.

Since 24/7/16 I have taken 50mg Nitrofurantoin each day at about 8pm to allow time for it to take effect before moving from wheelchair to bed (point

and pulled my catheter forward away from the bladder wall immediately after going to bed (point 11). From 24/7/16 - 12/8/16 I had no blockage in 20 nights.

So WITHOUT Nitrofurantoin and catheter repositioning: 10 nights, 6 blockages;

WITH Nitrofurantoin and catheter repositioning: 20 nights, no blockage.

It looks obvious that there is a connection between Nitrofurantion/catheter repositioning and stopping blockages. The strength of the evidence can be measured like this:

what are the chances of having 6 blockages on the first 10 nights (without Nitrofurantoin) and none on the following 20 (with Nitrofurantoin) if Nitrofurantoin and catheter repositioning had no effect?

That can be calculated like this:

if there had been just one blockage, the chances of it being in the first 10 nights would have been 10 divided by the total number of nights (30) = 0.3333;

if there had been two blockages the chances of both of them being in the first 10 nights would have been 0.3333 multiplied by 0.3333 = 0.1111 and so on ...... until

the chances of 6 blockages all being in the first 10 nights is 0.3333 multiplied by itself five times or 0.33336 = 0.0014, that is 14 chances in 10,000 or 1 chance in 714 (10,000 is the number that 0.0014 would have to be multiplied by to get 14). So the probability of there being no association between taking Nitrofurantoin/catheter repositioning is low, therefore the probability that there is an association is high. It is 1 - p = 0.9986 = 9,986 chances in 10,000 = 9,986/(10,000 - 9,986) chances in 10,000/(10,000-9,986) = 713 chances in 714.

.

The probability of there being no association between taking Nitrofurantoin/catheter repositioning and the occurrence of catheter blockages can also be calculated using the binomial distribution where the probability of success for a single trial is again 10/30 = 0.3333, the number of trials is 6 (the number of blockages on all 30 nights) and the number of successes (hardly the right word for a catheter blockage - the number of blockages during the first 10 nights) is also 6.

http://stattrek.com/online-calculator/binomial.aspxThe probability just calculated is 'if there is no association between Nitrofurantoin and catheter blockages, what are the chances of getting 6 blockages in the 10 nights without Nitrofurantoin?' It can also be calculated the other way round: what are the chances of there being no blockage in 20 nights with Nitrofurantoin if lack of blockages is not associated with it? In this case, the probability of success for a single trial is 20/30 = 0.6667 the number of trials is 6 (the number of blockages on all 30 nights) and the number of successes (the number of blockages on the last 20 nights) is 0. The answer is the same, that is, probability = 0.0014.

The multinomial distribution can also be used with 2 outcomes (blockage/ no blockage) probability of outcome 1 (blockage) is 10/30 = 0.3333; frequency of outcome 1 (number of blockages when not taking Nitrofurantoin) = 6; probability of outcome 2 (no blockage) is 20/30 = 0.6667; frequency of outcome 2 is 0 (the number of blockages when taking Nitrofurantoin). The answer is the same as before (p = 0.0014).

http://stattrek.com/online-calculator/multinomial.aspxThis too can be calculated the other way round: outcome 1 (no blockage, probability 0.6667, frequency 0); outcome 2 (blockage, probability 0.3333, frequency 6) with the same result.

So for the 30 nights of the trial, it is possible to say that taking Nitrofurantoin and catheter repositioning were associated with stopping my catheter blockages with only a very small chance of being wrong (probability 0.0014) which is 1 chance in 714.

That was the situation on 13/8/16. It is changing every day. If I have a blockage, the chances of the statement being wrong will increase. If I do not have a blockage, the chances of the statement being wrong will become even smaller; for example when the number of nights without a blockage reached 50 (12th September), the probability became (10/60) to the power 6 = 0.1667 to the power 6 = 0.000021, that is 21 chances in a million or 1 in 47,619 (one million is the number that 0.000021 would have to be multiplied by to get 21).

Such probabilities as p = 0.0014 or p = 0.000021 do not predict the frequency with which I can expect a blockage; neither do they predict the proportion of patients with a condition the same as mine having their blockages stopped. All they mean is that in my case, it is almost certain that taking Nitrofurantoin and repositioning my catheter have reduced the chances of getting a blockage: p = 0.000021 (the probability of there not being an association) or 1-p = 0.999979 (the probability of there being an association) are measures of the chance that Nitrofurantoin and catheter repositioning are associated with catheter blocking; they are not measures of what that association is, that is, how much they reduce the chances of a blockage.

So for patients with a condition the same as mine, the treatment can be expected to reduce the frequency of blockages for almost all of them.

It could be argued that although what has been calculated show a very strong correlation between taking Nitrofurantoin/catheter repositioning and the absence of blockages, demonstrating correlation is not the same as demonstrating cause: something else might have happened on 24th July 2016 when I started taking Nitrofurantoin regularly and it was this 'something else' which stopped my blockages. This is true, but I am not aware of anything happening on that day or soon after, and which has continued to be effective since then, which could be this 'something else'.

It is possible to predict the number of blockages within any specified period using the Poisson probability distribution:

http://stattrek.com/online-calculator/poisson.aspx.

For example, suppose we wish to predict the chance of 1 blockage in a period of 7 days, the Poisson random variable would be 1. The average rate of success is the average number of blockages which in the past have occurred in 7 days (number of blockages/number of days in observation period x 7). To predict the chance of 2 blockages in 28 days, the Poisson random variable would be 2 and the average rate of success would be the average number of blockages in 28 days. As long as there are no blockages on nights following taking Nitrofurantoin, the average number of blockages for any period is 0: so until there is a blockage, the prediction of future blockages for any period is zero.

NITROFURANTOIN AND BLOCKAGES 16th January 2016 to 16th January 2017

My first blockage occurred on the night of 16/17th January 2016.

There were 23 blockages in the first 190 days, none in the following 176.

If Nitrofurantoin and catheter repositioning had no effect, the probability of having 23 blockages in 190 days followed by no blockage in 176 days is (190/(190+176)) to the power 23 or 0.5191 to the power 23 = 0.0000002823 which is less than 3 chances in 10 million (ten million is the number that 0.0000002823 would have to be multiplied by to get 2.823).

http://www.rapidtables.com/calc/math/Exponent_Calculator.htmThe base is 0.5191 and the exponent is 23.

The probability of Nitrofurantoin having had no effect is becoming even smaller every day I do not have a blockage.

The binomial distribution can also be used with the same result:

http://www.vassarstats.net/textbook/ch5apx.htmlwhere N = 23, k = 23, p =0.5191 and the answer is p(k out of N)

The multinomial can be used too:

https://www.easycalculation.com/statistics/multinomial-distribution.phpwhere the number of outcomes is 2 (night/day), the number of occurrences 23 and 0, p = 0.5191 and 1- 0.5191 = 0.4809.

.

....... continued as part 2