Heart attacks in NZ – are women getting a raw deal?

Yesterday the NZ Herald published an article saying doctors are failing to spot heart attacks in thousands of women.  This sounds alarming, could it be happening in NZ? Are women getting a raw deal?  Important questions.  This post looks at the study behind the media and then at how heart attacks are being diagnosed in New Zealand.

The Herald article is evidently based on press releases related to a study published to coincide with the European Society of Cardiologist’s conference currently underway in Rome and attended by some 30,000 cardiologists, other physicians, industry types, and the Pope (yep!). The study in question comes from the University of Leeds. Here’s the Leeds Uni press release.

The US red-dress logo which is their national symbol for women and heart disease

The US red-dress logo which is their national symbol for women and heart disease

The study

The study is based on an audit of UK data collected about patients between 2004 and 2013 whose final diagnosis was a heart attack (in clinical jargon a myocardial infarction, either STEMI or NSTEMI).  The full article is available here.

The subjects were patients who at discharge from hospital had a heart attack diagnosis.  The authors looked at the preliminary diagnosis of patients when they first entered the hospital and compared that diagnosis to the final diagnosis of a heart attack.  The preliminary diagnosis was for about a 30% of patients not explicitly a heart attack – ie often something like “chest pain of uncertain cause.”  In the press release and news reports this was reported as a “misdiagnosis.”

Point 1:  The term “misdiagnosis” is inappropriately applied here.  While some forms of heart-attacks can be diagnosed in the ED, most can not.  Indeed, the guidelines for diagnosis of a heart attack require some blood measurements at least 6 hours apart.  Nowadays, the later blood sample is not done in the ED, but in a cardiology or general medicine ward.  That is, the ED physicians often don’t have all the data to make a definitive diagnosis – hence only a preliminary diagnosis is made. Most of the time the job of the Emergency department physicians is to rule-out some possible diagnoses and to identify patients at significant risk of a heart attack.  These patients are referred on to the specialist teams within the hospital who make the final diagnosis.*  Yesterday I was speaking with a cardiologist who was explaining how often cardiologists themselves disagreed over a diagnosis.  It ain’t easy.

The press releases and media reports emphasise that a larger proportion of women than men were likely to have a change between the preliminary and final diagnosis.  The Leeds University press release states women were 59% for a final STEMI diagnosis and 41% for a final NSTEMI diagnosis more likely than men to have a change from the preliminary diagnosis.

Point 2:  These numbers are not reported in the published paper!  Nor is anything about the differences between men and women discussed in the paper.  In the results section it is simply stated that those who had an initial diagnosis that changed were more likely to be older, female, and have a co-morbidity. There are some numbers related to this in a table. In the table I note that patients older than 61 compared with younger patients had at least (more if they were even older) the same odds of having a diagnosis changed as did females compared with males (it’s a little awkward in the paper because the odds ratio is written the opposite way around – but this can be rectified simply by taking the reciprocal of the odds ratio and comparing that).  There were also other predictors of a change in diagnosis (eg higher heart rate).  The cynic in me thinks that it may be for publicity reasons that the emphasis has been placed on the sex differences in press releases.

Point 3: What is important about the study is that in those who had a change in diagnosis the one-year mortality rates were higher.  While the suggestion is made that this is because of delay in time to treatment (known from other studies to be important), there are other potential reasons because of the differences in demographics and co-morbidities between the groups.

New Zealand

The study began at a time when the blood biomarkers indicative of a heart attack that are used now (troponins) were not in common use.  There have been several generations of markers, the latest of which are “high-sensitivity troponins.”  The authors’ recommended that:

“…our results… call for the earlier use and wider adoption of high sensitivity troponins as well as a focus on the systematic application of accelerated diagnostic protocols using risk scores rather than subjective clinical assessment.”

The good news is that New Zealand is now the only country in the world** to have accelerated diagnostic protocols using risk scores in place in every ED.  Furthermore, most ED’s are using the latest high sensitivity troponins.

In the Christchurch ED, different sex-specific thresholds of the troponin used for risk stratifying and diagnosing heart attacks are used.  This is because in the general healthy population males have slightly higher values of these troponin measurements than females.  Therefore, to avoid underdiagnoses of females a lower diagnostic threshold is used.  Furthermore, in a study we were part of and lead by our Brisbane based colleagues, using sex-specific threshold helped improvs risk prediction for future adverse events in women.


In New Zealand it is less likely that women are getting a raw deal.


*perhaps the Pope a.k.a @Pontifix [literally the “bridge builder”] could help bridge the divide between ED physicians and cardiologists – generally ED physicians rule-out heart attacks, Cardiologists rule-in heart attacks.

** although Queensland also has this and they like to think of themselves as a country sometimes

Cheesecake Files: Embargoed until

Every now and again a Journal doesn’t want us to talk publically about our own paper until it they publish it.  This is simply so they can make more of a splash with it.  This was the case of an article I have been involved with published today in the Cardiology journal of the American Medical Association*. 

What’s it about?

Ruling out a heart attack in the emergency department is difficult.  Readers of this blog would have read about various other ways we’ve developed to be part of it (eg here).  They depend on many things including the type of blood measurement used and the timing of that.  These markers – called troponins – detect damage to the heart muscle.

In this study led by Ed Carlton of Southmead hospital in Bristol, UK, we evaluated whether a single measurement of very low levels of a comparatively new blood biomarker called “High-sensitivity troponin I” could alone rule out heart attacks within 30 days when someone presented to the emergency department with chest pain (Here Ed speak about the study here).

3155 patients at 5 hospitals in New Zealand (Christchurch), Australia and the UK participated of whom 291 had a heart attack (277 were having a heart attack on the day they presented to ED, the other 14 had one within 30 days).

We found that 594 (18.8%) of patients had such a high sensitivity troponin I concentration below the limit at which it could reliably be detected.  ie next to nothing. These we can say tested negative.  Three of them (0.5%) it turned out did have a heart attack.

Why the splash?

The editor got quite excited about this and another study and wrote an editorial to accompany the studies:

“To manage costs and the adverse effects of overcrowding in the ED, it is a high priority to rapidly and safely identify patients with a sufficiently low probability for acute coronary syndrome (<0.5%-1%) so that they can be discharged efficiently and avoid unnecessary testing.”

He was impressed that this study had been “tested in robustly sized, geographically diverse, clinically relevant populations.”

And concludes

“Taken together with prior studies the findings from the studies of Neumann et al and Carlton et al lend strong support to the notion that accelerated diagnostic protocols that incorporate [high sensitivity troponins] can facilitate earlier triage while maintaining an acceptable [rate of false negatives].

One of the features of interest to the readers of this US journal is that the high-sensitivity troponins are not yet available in the US, however they eventually will be and how they are used is of particular interest.

Part of a figure from the publication showing how choosing different troponin thresholds to rule-out patients affects how sensitive the test is.

Part of a figure from the publication showing how choosing different troponin thresholds to rule-out patients changes how sensitive the test is.

Why not perfection?

Of course we would love to never have a false negative (or false positive).  However, the reality of medicine is that this is not possible.  We could, of course, simply admit everyone, do more invasive tests,  or “wait and see” if they develop a heart attack. There are, though, risks as well as costs with admitting people to hospitals.  If I may speculate for a moment, the rise in superbugs resistant to antibiotics is only likely to increase those risks in the future – hence the importance of studies such as this. It is important that we get the balance of risk right.

What are the next steps?

All of New Zealand now has some kind of accelerated diagnostic pathway for chest pain patients that incorporates serial troponin measurements.  At some stage we will implement, monitor, and measure the addition of an even more accelerated rule-out for some patients with just one troponin.  Watch this space.


* Carlton, E., Greenslade, J., Cullen, L., Body, R., Than, M. P., Pickering, J. W., Aldous, S.A., Carley, S,. Hammett, C., Kendall, J., Kevill, B., Lord, S.J., Parsonage, W.A., Greaves, K. (2016). Low concentrations of high-sensitivity cardiac troponin I at presentation in the evaluation of emergency department patients with suspected acute coronary syndrome. JAMA Cardiology. http://doi.org/10.1001/jamacardio.2016.1309

World Kidney Day 2016

Cheesecake files: A little something for World Kidney Day

Today is World Kidney Day, so I shall let you in on a little secret. There is a new tool for predicting if a transplant is going to be problematic to get working properly.

Nephrologist call a transplant a “graft” and when the new kidney is not really filtering as well as hoped after a week they call it “Delayed Graft Function.”  Rather than waiting a week, the nephrologist would like to know in the first few hours after the transplant if the new kidney is going to be one of these “problematic” transplants or not.  A lot of money has been spent on developing some fancy new biomarkers (urinary) and they may well have their place.  At this stage none are terribly good at predicting delayed graft function.

A while ago I helped develop a new tool – simply the ratio of  a measurement of the rate at which a particular substance is being peed out of the body  to an estimate how much the body is is producing in the first place.  If the ratio is 1 then the kidney is in a steady state. If not, then either the kidneys are not performing well (ie not keeping up with the production), or they have improved enough after a problem and are getting rid of the “excess” of the substance from the body.  This ratio is simple and easy to calculate and doesn’t require extra expense or specialist equipment.

A few months ago, I persuaded a colleague in Australia to check if this ratio could be used soon after transplant to predict delayed graft function. As it turns out in the small study we ran that it can, and that it adds value to a risk prediction model based on the normal stuff nephrologists measure! I’m quite chuffed about this.  Sometimes, the simple works.  Maybe something will become of it and ultimately some transplants will work better and others will not fail.  Anyway, it’s nice to bring a measure of hope on World Kidney Day.

This was published a couple of weeks ago in the journal Nephron.



Cheesecake files: A world second for heart attacks

Going to the Emergency Department with chest pain no longer means an almost certain night in hospital.  Friday saw the publication online of our randomised controlled trial comparing two different strategies to rapidly rule-out heart attacks in people who present with chest pain to hospitals.  Here’s a précis:

What’s the problem?

  • Chest pain is common – 10% or so of presentations to ED are for chest pain.
  • Heart attacks are not so common – only ~10-15% in NZ (and less overseas*) actually have a heart attack.
  • It is devilishly difficult for most chest pain to rapidly rule out the possibility of a heart attack.
  • Consequently, most people get admitted to hospital (in 2007 93% of those presenting with chest pain).

But – led by Dr Martin Than in Christchurch and an international group including Dr Louise Cullen in Brisbane – a series of observational studies and one randomised control trial have resulted in a gradual increase in the proportion discharged.  The trial was the first of its kind, it compared standard practice at assessing chest pain to a purpose built accelerated diagnostic pathway (ADP), which we called ADAPT.   In that study 11% of patients in the standard practice (control) arm and 19.3% in the ADAPT ADP arm (experimental arm) were discharged home from ED within 6 hours.  A great improvement which led to that ADP being adopted in Christchurch hospital.

So why another study?

Two reasons: First, 19% still means that there are many patients being admitted who potentially don’t need to be in hospital.  Second, the ADP was based around a risk assessment tool designed to rule-in heart attacks rather than rule-out.  In the meantime, the team had constructed a purpose build risk assessment tool that in observational studies looked like it could rule out 40-50% of patients.

What is the study just published?

The world’s second randomised controlled trial of assessment of chest pain compared the ADAPT ADP in use (now the control arm) with a new ADP based on the new Emergency Department Assessment of Chest pain Score (EDACS)[the experimental arm].  The only difference between the two arms of the study was the risk assessment tool used. The tool gave a risk score. Patients with a low score, no unusual electrical activity in the heart, and no elevated heart muscle injury proteins in either of two blood samples measured 2 hours apart, were considered low risk.

An important aspect of the study was that it was pragmatic.  This means that the doctors didn’t have to follow the ADP and could decide to send a patient home, or not send them home, based on any factors they thought clinically relevant.  This makes it very tough to run a trial, but it makes the trial more “real life.”

What were the results?

558 patients were recruited.  They all volunteered and are marvellous people.  I love volunteers.

The primary outcome was the proportion of patients safely discharged home within 6 hours.  We assessed safety by looking at all medical events that happened to a patients over 30 days to check to see if any patients discharged home had a major cardiac event that could potentially have been picked up in the ED.

34% of the control arm and 32% of the experimental arm were discharged within 6 hours.  In other words, there was no difference in early discharge rates between the two arms.  The surprising feature of this is that between 2012/3 (when the first trial was run) and 2014/15 the proportion of patients the first ADP ruled out increased from 19% to 34%.  This was unexpected, but pleasing. There were no safety concerns with any patients.

The secondary outcome was simply the proportion each arm of the study classified as low risk (ie not considering whether this led to early discharge or not).  The control (ADAPT ADP) classified 31% and the experimental (EDACS ADP) 42%.  This was a real and meaningful difference which suggests that there is “room for improvement” in early discharge rates as the clinicians become more familiar with the EDACS ADP.

Since 2007 in Christchurch hospital over three times more patients who present with chest pain can be reassured from within the ED that they are not having a heart attack and discharged home (see the infographic).EDACS infographic v2

What was your role?

My role: I managed aspects of the data collection for the later 2/3rds of the patients recruited, did the statistical analysis and co-wrote the manuscript.  In reality, there were a lot of people involved, not least of whom were the wonderful research nurses and database manager who did a lot of the “grunt work”.

What now?

Over the last year all EDs in New Zealand have implemented or in the process of implementing an accelerated diagnostic pathway.  The majority have chosen to use the EDACS pathway.  I am part of a team nationwide helping implement these pathways and monitor their efficacy and safety.


This study was funded by the Health Research Committee of New Zealand. The work was carried out with the collaboration or the University of Otago Christchurch, Christchurch Heart Institute, and the Canterbury District Health Board Emergency Department, Cardiology Department, General Medicine, and Canterbury Health Laboratories. My salary is provided through a Senior Research Fellowship in Acute Care funded Canterbury Medical Research Foundation, Canterbury District Health Board and the Emergency Care Foundation.

*Not because we have more heart attacks, just an efficient and well funded primary care sector that keeps the very low risk patients out of the ED.

**Featured Image: Creative Commons Share-Alike 3.0 http://tcsmoking.wikispaces.com/heart%20attack



The effect of cannabis legalisation on Emergency Care 

Medical cannabis application guidelines are to be reviewed, announced Associate Health Minister Peter Dunne this week. Co-incidently a paper was published* in the Annals of Emergency Medicine on the effect on Emergency Care of legalising medical marijuana use in Colorado. Alas, this article is behind a paywall.  It does not have a lot of detail. However, it is relevant to the New Zealand debate.  Not so much as to any possible change in guidelines on applications made to the minister, but rather to the effect a broader legalisation of marijuana for medical purposes may  have an on our emergency departments. i.e. just one of the many factors which need to be taken into account in the debate.

In 2009 the prosecution of marijuana users and suppliers was halted in the state of Colarado where the use of medical marijuana had been previously legalised and licensed.  Within 2 years the number of registered medical marijuana licences increased 24 times from 5000 to nearly 120,000.  This was not the only effect:

  • The percentage of 18 to 25 year olds reporting marijuana use increased from 35% to 43%
  • The percentage of those aged 26+ perceiving marijuana posed “great risk” dropped from 45% to 31%.

While these numbers may reflect in part the readiness to be “honest” after the law change, the following statistic probably is truly related to increased use:

  • The hospitalisations after marijuana use nearly doubled from 15 per 100,000 hospitalisations to 28 per 100,000 hospitalisations.

As the authors concluded:

“It is clear that marijuana availability and use in Colorado significantly increased after the commercialization of medical marijuana. Providers in states with impending legalization measures should become familiar with the symptoms and management of acute marijuana intoxication, as well as understand the effects on chronic diseases frequently observed in the ED.”

I was fascinated that in a population of 5.4 million there were nearly 120,000 licensed medical marijuana users in Colorado.  That is 1 in 45 people! That strikes me as amazingly high proportion. However, I guess that it all depends on just what medicinal purposes may mean.  The process to get a license (at least now), seems relatively easy involving a few simple forms.  The Physician recommends the number of plants to be grown and ticks a box stating one of eight conditions: Cancer, Glaucoma, HIV or AIDS positive, Cachexia, Severe Nausea, Severe Pain, Persistent muscle spasms, Seizures.  The patient sends in a form too, with $15.

According to the latest statistics on the Colorado medical marijuana registry there are currently 107,000 active patients registered with an average age of 43. 21-40 year olds comprised 43% pf the patients.  93% report severe pain.  Hmmm… it seem Colorado has an epidemic of “severe” pain amongst their young adults.


ps. Before you jump in with comments, recognise that there is a lot of misconception around medical cannabis in New Zealand. Minister Peter Dunne cleared some of them up in a press release in January. (eg did you know that there is already a cannabis product approved for therapeutic use?).

Note: Recreational marijuana became legal in Colorado in 2014.

Disclaimer:  I an not an expert in the field, merely I came across this article because it was published in a journal I access for my other research concerning emergency departments.  If you believe the methods to measure these things, Ann Emerg Med is the top ranked Emergency medicine journal

*Kim, H., & Monte, A. A. (2016). Colorado Cannabis Legalization and Its Effect on Emergency Care. Annals of Emergency Medicine, http://doi.org/10.1016/j.annemergmed.2016.01.004

Photo: Public Domain, from Wikipedia.

What’s going on at the UOC?

Q. What has Mars, Epidemics, Heart Disease, Infection, and Pacifika got in common?

A. They are all central to research project at the University of Otago Christchurch (UOC).

Here are some excerpts for the UOC summer newsletter (Written by UOC communications manager, Kim Thomas).

Christchurch in NASA Mars project role

University of Otago, Christchurch, researchers are playing a crucial role in research that will assist in NASA’s mission to Mars.

Screen Shot 2016-02-09 at 10.15.21Thee Christchurch researchers are scanning the brains of explorers who have wintered in Antarctica as part of a NASA /German Aerospace Center project to understand what impact living in extreme environments has on the human brain. The research will be relevant for NASA’s plans to send humans to Mars. The shortest possible return trip to the red planet would take two years.

The international research team is led by the University of Pennsylvania’s Associate Professor Mathias Basner. His team will be scanning the brains of astronauts, while the Canterbury team focuses on those who have wintered in Antarctic’s extreme and isolated environment.

Dr Tracy Melzer is the MRI research manager for the Christchurch campus’ New Zealand Brain Research Institute. He says the research aims to understand whether prolonged periods in these extreme, isolated and hostile environments change brain structure and function.

His international collaborators have already found the hippocampus region of the brain, which is important for memory formation and visual/spatial orientation, actually shrinks during the Antarctica winter.

Dr Melzer and his colleagues will scan the brains of up to 28 international explorers over two years. They are tested before leaving for Antarctica, immediately on their return, then six months afterwards. The Christchurch scans are important because they capture explorers immediately as they return from the ice.

Preparing for future disease epidemics

Christchurch microbiologist Professor David Murdoch has taken part in an invitation-only global think tank aimed at better anticipating future infectious disease epidemics.

The head of the University of Otago, Christchurch’s Pathology Department was one of two Australasians invited to the World Health Organization-led event late last year.

Professor Murdoch says he was privileged to be among about 130 international experts invited to attend, including human and animal health experts, and members of aid agencies and the insurance and travel industries.

“ The big idea was how to better prepare for future epidemics, knowing there definitely will be ones. It also recognized reviews of the Ebola response and a desire to improve on that.”

Acknowledging the importance of collaboration, one key outcome of the event was getting people from diverse areas of expertise together, Professor Murdoch says.

Thee event consisted of six sessions, including ‘Back to the future: learning from the past’, and ‘Preventing the spread of infectious disease in a global village’. Each session consisted of short talks by five experts, then robust discussion.

Professor Murdoch spoke at the event about the relatively new area of microbiomes (the communities of microorganisms that inhabit parts of the human body) and how understanding it could help with preparing for and controlling future respiratory disease epidemics.

Some of the ideas that emerged from the event were that global and public health were getting more political attention than ever, and that health threats increasingly reflected nature, including the animal world, and so acknowledging and understanding its interplay with human health was important.

Contact between children monitored in world first infection study

Christchurch primary school pupils are wearing sensors tracking contact with each other in a world-leading study to better understand a common but serious disease.

The staphylococcus bacterium is a major cause of serious infections such as septicaemia, but also often presents as sores on the skin. Most commonly, though, it is carried harmlessly on skin or in noses, from where it can be passed on to others who might become ill. Very little is known about who passes it to whom in the community.

University of Otago, Christchurch researcher Dr Pippa Scott is testing levels of the bacteria in Linwood Avenue School pupils and, in a world first, monitoring contact between them using ‘proximity sensors’ to better understand how staphylococcus is passed from person to person.

Dr Scott says school-aged children o en spread u and other diseases so could be important to the spread of staphylococcus in the community.

“We asked a lot of schools if they would take part in the study and Linwood Avenue School principal Gerard Direen came back to us quickly and said the school would be really keen to help.’’

Dr Scott says 70 children aged between 8 and 11 were given the proximity sensors to wear clipped to their shirts for around 2 weeks. e sensors are not GPS devices and cannot pinpoint a child’s whereabouts but rather record when children come in contact with each other. They have never before been successfully used in a study linking infectious disease spread to contact in the same individuals.

The study is ongoing but early analysis found almost every child was carrying the bacterium at some stage during the seven times they were tested. More than half the children carried the bacteria at any one test session. Almost all strains the children had were susceptible to commonly prescribed drugs for the condition.

First study of South Island Pasifika heart health

“She was one of the first scientists to demonstrate our cells produce free radicals as part of their normal function.”

It’s well known that New Zealand’s Pacific population suffers higher rates of heart disease than the general population. But until now, evidence has been based on data gathered
in Auckland. University of Otago, Christchurch researcher Dr Allamanda Faatoese is changing that with the launch of the Pasifika Heart study of Christchurch Pacific people.

“Pacific communities living in Auckland have vastly di erence environments than those in Christchurch. We know little about the heart health pro le of Pasifika people in Christchurch,’’ she says.

The Heart Foundation-funded Pasifika Heart study will for the first time measure heart disease risk factors in 200 Pacific Island participants, both healthy people and those su ering from illness. Dr Faatoese is based at the University’s Christchurch Heart Institute but will study participants from across the South Island.

Each participant’s personal and family medical history, blood pressure and body composition will be recorded along with their cholesterol levels, blood sugars and markers linked with kidney function, gout and heart failure.

(c) Creative Commons. Intangible Arts https://www.flickr.com/photos/intangible/

The physics of maiming a child (repost because of another close encounter)

Dear Driver,

When you backed out of a driveway and did not even see how I swerved around behind your car to avoid T-boning you, how dare you have the temerity to tell me you were careful!  I was 7 feet tall, dressed in bright yellow and traveling at no more than 10 km/h.  Perhaps a simple lesson in physics will help you and your fellow “driveway backers” to realise how dangerous you are and to adopt safer driving practices.

In the diagram you can see a car backing out of a driveway.  Typically when you are at the edge of your property and have a fence (see photo below) blocking your view of the footpath you are able to see about 1.7 metres along the footpath.  Let us imagine that there is a child on a trike riding at 5 km/h just out of your line of sight.  How long  does it take them to travel that 1.67 metres?  The physics is quite easy.

Car backing out of a driveway. Illustration of how little of the footpath can be observed.

Car backing out of a driveway. Illustration of how little of the footpath can be observed.

Velocity = distance/time, therefore time = distance/Velocity.

5 km/h is 5000 metres in 60 x 60 seconds, ie about 1.4 m/s.  Putting this in the formula above means that it takes about 1.2 seconds for the child to travel that 1.67 metres.

Now consider this. According to design guidelines for safe bicycle use 2.5 seconds must be allowed for someone to observe the danger, react, apply brakes and stop.  In other words, if you covered the distance from your driveway to the middle of the footpath, about 1 metre, in under 1.2 seconds you will almost certainly hit the child.  That is a speed of just 3 km/h!!!!!

Now consider who else is on the footpath, all legally:

  • Pedestrians 5 km/h
  • Joggers 5- 15 km/h
  • Kids on skateboards or scooters 10 km/h
  • Child on bicycle with small wheels, 10 km/h
  • Mobility scooter, 5-10 km/h
  • Me on my Trikke, 10 km/h
  • Postie on a bike 5-10 km/h.

For those going 10 km/h your speed needs to be just over 1.5 km/h to hit someone!

So, before you do some damage here is what you can do:

  • Never back out of a driveway unless you really really must.  If you think you must because of the design of your driveway, change the design!
  • Cut back those hedges, remove some of that fence so that you can see further.
  • Always always always stop at the end of your driveway (BEFORE THE FOOTPATH) and toot a horn.  Then proceed very very slowly.

By the way, you are legally obliged to give way:

 Land Transport (Road User) Rule 2004

4.4 Giving way when entering or exiting driveway

A driver entering or exiting a driveway must give way to a road user on a footpath, cycle path, or shared path (as described by clause 11.1A(1)).

Thank you for considering the physics of maiming a child, may you never find your self in such a terrible situation.


Dr John Pickering

A typical driveway with almost non-existant visibility

A typical driveway with almost non-existent visibility


Feature Image: Intangible Arts https://www.flickr.com/photos/intangible/ under Creative Commons Attribution 2.0 licence.